Demographic change and assimilation in the early 21st-century United States

Significance

The contemporary intellectual landscape, dominated by theories about race and racism, has engendered the majority–minority conception of the American future, in which White people are outnumbered by Americans of color by midcentury. But assimilation, as a set of processes with societal ramifications, is still potent, driven by demographic dynamics that generate opportunities for minority mobility, and is linked to increasing family mixing between White and non-White people. The children of these families offer the best window into ongoing assimilation. The binary vision of the majority–minority society needs major modification because of the emergence and growing size of this bridging group. In addition, racism-focused theories need to be complemented by assimilation ideas to grasp the implications of demographic change.

Abstract

Assimilation has not gotten its due in recent decades, overshadowed by discussions about race and racism. Using evidence from the 2000 Census and the 2015 to 2019 American Community Survey, we show, however, that contemporary social change and its implications for a society stratified along ethnoracial lines cannot be understood without taking the assimilation of many Americans with non-White or Hispanic family origins into account. We begin by necessity with a reconsideration of the concept, arguing that assimilation involves the decline in influence of ethnoracial origins on social status and on relationships with others; it does not require the erasure of all markers of those origins. We highlight that demographic dynamics are playing a critical role in promoting assimilation: Ethnoracial shifts across birth cohorts are generating opportunities for minority mobility, and increasing mixing in families is engendering a growing group of young Americans from mixed minority–White family backgrounds. These mixed Americans, we argue, offer the clearest window into assimilation processes, which are not limited to them, however. The empirical evidence presented shows effects on both social status and relations across group boundaries: 1) the top tier of the workforce, the best-paid quartile of occupations, is growing much more ethnoracially diverse over time, and mixed minority–White Americans are important to this diversity because their probabilities of entering this tier are more similar to those of White people than are those of unmixed minorities; and 2) intermarriage remains robust, enhanced by high rates of marriage to White people by mixed minority–White individuals.

Ideas about race, racism, and racialization today are dominating both the social-scientific and popular discussion of majority–minority relations. (The terms “minority” and “majority” refer by definition to the positioning of social groups within a societal hierarchy of status and power, but in the United States also correspond with their current numerical size.) This paper will argue, however, that to understand the changes taking place in American society in the early 21st century, we need assimilation ideas, too.

Race-focused theories, such as racial-formation and critical-race theories, have demonstrated their value in revealing racial inequality as a persistent feature of American society and in illuminating its linkages to slavery and conquest in the nation’s history and to the institutional mechanisms that maintain it in the present (1–3). Race theories generally assume that the boundaries separating racial groups are rigid and impose inequality, although they may be occasionally penetrated by individuals who are “passing” from one group into another. A race lens produces a binary view of American society as divided between White people and “People of Color.” This view in turn implies a majority–minority future in which the current majority, White people, becomes a numerical minority because of demographic change (4).

In contrast, assimilation theory is attentive to social processes that weaken group boundaries and their social significance, at least for parts of groups (5). It holds that these boundary changes are associated with processes of social mobility for minorities that bring many of them into equal and socially more intimate contact with members of the dominant group. In principle, this holds even when racial boundaries are involved, although assimilation scholars generally recognize that the White/Black boundary has been stubbornly resistant to major change. There is abundant evidence of assimilation in the American past.

Because of assimilation’s association with groups now deemed “White,” race-focused scholars usually reject a contemporary role for assimilation resembling its past one (3, 6, 7). A compelling reason, however, for considering such a role is assimilation’s historical availability to immigrant-origin groups (as compared to groups incorporated through enslavement and conquest). During the second half of the 20th century, assimilation made it possible for Catholic and Jewish descendants of late 19th- and early 20th-century immigrants from mainly southern and eastern Europe to become full members of the societal mainstream, despite the exclusion they once confronted.

Today, much of the society’s demographic dynamism is again the byproduct of immigration, associated not only with immigrants but with their descendants, who make up an increasing share of young Americans. Because of their primarily non-European origins, immigrants and their children occupy an ambiguous position with respect to these bodies of ideas. As non-White people (mostly), many are affected by institutional and individual racism (1, 2, 8), but many are also assimilating, as they aspire to social mobility, which at this historical moment is most often connected with integration into social milieus containing many White people (5). Moreover, immigrants and their children are especially prone to believe that they can overcome discrimination through additional effort, and they appear in general to be viewed more favorably by White people than are African Americans (9–11). In addition, the selectivity of immigration means that some groups, especially Asian ones, arrive with high levels of human capital, primed for mobility (12). Rising intermarriage and, in the next generation, mixed minority–White family backgrounds provide powerful evidence of the integration of many with White people (13).

In this paper, we will demonstrate an important contemporary role for assimilation through empirical evidence of its patterns. In so doing, we will extend an earlier analysis (14) by demonstrating the central role of ethnoracially mixed individuals in the growing diversity at the top of the US occupational hierarchy and presenting new intermarriage data that reveal an important change for Americans of mixed Black and White origins. We will also present a population pyramid confirming the strong differentiation of the mixed population by age. But, comprehending the full import of this evidence requires revisiting and revising widespread ideas about assimilation; we start there.

Assimilation Ideas for the 21st Century

A reworking of a widespread conceptualization of assimilation is necessary to grasp the concept’s contemporary relevance. The sociologist Milton Gordon developed what has long been accepted as the canonical formulation of assimilation (15). At its core, this formulation envisioned a one-way process by which a minority group came to resemble the majority in its culture and social status and was eventually absorbed by it, taking on its identity.

Written as the massive post–World War II assimilation of the descendants of Catholic and Jewish European immigrants was underway, this formulation failed to fully reflect the eventual outcome. Pointedly, the mainstream had been dominated until then by an exclusively White Christian (Protestant) majority, which maintained its dominance in part by virulent prejudice and discrimination, including at times violence, directed toward the immigrants and their descendants (16). But the mainstream underwent a major expansion as a result of assimilation processes (17). One indicator is the shift of its religious identity to Judaeo-Christian (18). (The Google Books Ngram Viewer shows that the use of this term increased about 10-fold between the late 1930s and 1960.) Many of the newly assimilating children and grandchildren of immigrants also asserted previously marginal ethnic or hyphenated identities, albeit most often in symbolic forms that involved limited commitments in everyday life (19–21), thereby creating visible ethnic diversity in the mainstream. Its culture—visible in television, cinema, and novels—expanded, as it increasingly reflected the personae and experiences associated with the new groups. Ethnoreligious distinctions among White people have waned to the point that today, for the first time in American history, all three branches of government are headed by observant Catholics (President Joe Biden, House Speaker Nancy Pelosi, and Chief Justice John Roberts), at the same time as the leader of the Senate is an observant Jew (Charles Schumer), a configuration that was unthinkable and deeply objectionable to many White Protestants as late as the Kennedy election of 1960 (22).

In light of this history, Alba and Nee (5) reformulated the concept of assimilation to recognize its possibly two-sided nature. Their definition—“the decline of an ethnic distinction and its corollary cultural and social differences”—implicitly acknowledges that change can occur on both minority and majority sides (5). It envisions assimilation as a set of processes that, for some members of a minority or for the group as a whole, reduce the influence of ethnoracial origins on social status and social relations in relation to the majority group. It is agnostic about the long-run survival of some distinctive traits and identities. The decline of a distinction—a perceived social difference rooted in ethnoracial origins—implies the potential for greater social intimacy between majority and minority.

This version of assimilation theory posits that a powerful mechanism driving assimilation is the motivation of many minority-group members to improve their economic and social situations. Since in general only modest improvements are possible within ethnoracial communal spheres, those seeking mobility usually are pushed toward the mainstream society. The mainstream by definition reflects the ethnoracial hierarchy of the larger society; that is, the institutions that mobile minorities must participate in to move ahead, such as education, embody in multiple respects the power of the majority (23).

The well-known and amply supported contact theory from social psychology (24) spells out some of the consequences of minority participation in mainstream institutions. In such contexts (e.g., classrooms, work sites, and neighborhoods) the minority and majority encounter one another in situations of formal equality (e.g., students in the same class), a condition that enhances the stereotype-reducing effects of contact. Under such circumstances, contact can “individualize” the member of another group, whether majority or minority, and reduce over time the salience of an ethnoracial membership (25). Contacts of this type are very different from the more distant interactions that encourage people to view each other solely in terms of social categories (26).

Because assimilation as just described does not require the extinction of ethnic traits (i.e., distinctive identities and cultural features associated with an identity), Alba and Nee (5) argue that it is better conceptualized as entry into the mainstream rather than incorporation into an ethnoracial majority. Some assimilation may over time, marked in generations, result in disappearance into the majority. But for many, ethnoracial origins do not disappear but exert little influence in the everyday.

The theory just sketched identifies mechanisms of assimilation that in principle operate continuously. However, there is abundant evidence that the magnitude of assimilation—in the sense of the proportions of minority groups affected by it—is greater at some moments than at others. This temporal variation depends on larger conditions, such as the state of the economy, local or societal or, as is currently the case, demographic ones.

Thus, the large-scale assimilation of the post–World War II period was fostered by the economic prosperity of the postwar decades. It in turn led to a massive expansion of postsecondary education, especially of public institutions; the number of students in colleges and universities grew fivefold between 1940 and 1970 (5). And the occupational structure shifted in ways that created opportunities for the new graduates to find jobs appropriate to their status. Such expansions of opportunity are consequential in part because they reduce the likelihood that newly ascendant minority individuals, arriving in social contexts where they have been absent before, are seen mainly as competitors.

In the early 21st century, the societal condition that creates such non–zero-sum mobility (to be understood, at its most general, as the entry of ethnoracially marked newcomers into social spheres in which they have been previously absent or rare) is demographic (14, 27). Specifically, the largely White and well-placed baby-boom cohorts are leaving the ages of economic activity; the youthful cohorts entering them are more diverse and, in particular, contain substantially fewer White people. Therefore, there are not enough White people to replace those leaving, to say nothing of filling positions created by workforce expansion.

Measuring Race and Ethnicity

Pivotal for this analysis is the operationalization of race and ethnicity in a way that is revelatory of assimilatory change and, in particular, identifies as fully as possible those from mixed minority–White family backgrounds. The 2020 Census has called new attention to the growth of the mixed population (28), but our focus is on that part of it whose members are most likely to have grown up with a White parent and a minority one (and hence are “first-generation mixed”) (29). They are in a unique sociological position because they have parents and other kin belonging to both minority and White groups. In other words, they are individuals with minority ancestry who have access from birth to the White group and its cultural codes and perhaps to some of its advantages as the most powerful ethnoracial group.

The census measurement of race goes back to the first census in 1790, implied by the categories specified in the census clause of the Constitution (article 1, section 2). The categories have changed numerous times since then (30). Since 1980, the key questions have been twofold: one on race and the other on Hispanic origin (an ethnicity because the Hispanic category is not inherently racial). The questions, as they appeared in the 2019 American Community Survey, are shown in SI Appendix, Table S1.

Until 2000, the census classification of Americans into ethnoracial categories required these to be mutually exclusive: everyone must fit in just one. This requirement was bolstered by the popular belief in the “one-drop rule,” which held that mixed racial ancestry (especially Black and White ancestry) should be viewed as membership in the minority race. Starting in 2000, partly in response to the demands of the rising number of parents in mixed marriages, the census allowed multiple responses to the race question (30). However, the measurement of mixed ethnoracial backgrounds remained incomplete because 1) the reporting of mixed race for the same individuals is inconsistent across census datasets (14, 31); and 2) the two-question format question does not allow individuals to describe themselves as having mixed-Hispanic and non-Hispanic ancestry (14). That is, anyone who reports any Hispanic ancestry is regarded as only Hispanic in standard census classifications, regardless of their responses elsewhere.

To achieve a fuller depiction of the mixed minority–White part of the population, we draw on American Community Survey data, along with the Public Use Sample of the 2000 Census (32–34), because these datasets include a variable—ancestry—that did not appear in the complete-count decennial census questionnaire prior to 2020. (For the ancestry question, see SI Appendix, Table S1.) In brief, we classify individuals as mixed if they appear—from the combined information in the Hispanic origin, race, and ancestry variables—to have both White (or European) and non-White ancestries. That is, non-Hispanics are of mixed minority–White race if they report “White” and another race on the race question or if they report a single race on the race question and their response to the ancestry question indicates a different race (so long as one of the races is White). For example, a person who checks only “Black” on the race question and then indicates Irish ancestry would be classified in the mixed Black and White category. We classify a person as mixed Hispanic and White if she checks non-Hispanic and White by race but reports a Latin American ancestry, such as Mexican, or if he checks Hispanic and then reports a European ancestry (other than Spanish) (see SI Appendix, Tables S2 and S3 for precise category definitions). We underscore that this definition does not automatically classify persons who report as Hispanic and White by race (the way many Hispanics classify themselves racially) as part of the mixed population; the ancestry information for such individuals must indicate non-Hispanic White family origins for them to be classified here as mixed Hispanic–White.

In order to restrict the mixed minority–White group to likely first-generation mixed individuals, we do not include in it those who appear to have three or more distinct ancestral strands (e.g., American Indian–Black–White). (They are counted in an “other” category that plays little role in our analysis and also contains mixed individuals of exclusively minority origins.)

We also define a series of unmixed racial categories: Asian, Black, Native American (i.e., American Indian), and White; all of these categories are limited to non-Hispanics and exclude anyone whose ancestry responses imply mixture. Unmixed Hispanics are persons of any race who check Hispanic, so long as their ancestry does not indicate a non-Hispanic ancestor.

The Significance of Contemporary Demographic Conditions for Assimilation

The demographic dynamics promoting assimilation are evident in a population pyramid that shows the ethnoracial compositions of age groups, such as Fig. 1, constructed from 2019 American Community Survey data. This pyramid features an exhaustive division of the population among unmixed White people, mixed minority–White people, and minorities (who here include those in the “other” category).

Fig. 1.

Population pyramid by ethnoracial origin (White, mixed, and minority), 2019 Source: American Community Survey data (32, 34).

Visual inspection of the population pyramid reveals a sharp drop-off in the number of unmixed White people (colored red in Fig. 1) between the youngest of the baby boomers (born between 1955 and 1964 and aged 55 to 64 y in 2019), edging toward a decline in labor-market activity, and the youthful group entering the labor market (20 to 29 y old in 2019, for example). The fall-off is about 20%. And it continues into younger cohorts, still primarily in the ages of children. Compared to the youngest children of 2019 (0 to 9 y old), the decline in the number of White people from the baby-boom era is much more massive, about 36%. Even when we compare these children to a cohort that will be on the edge of retirement in two decades, when the children are entering the labor market, there is still an appreciable decline in the number of unmixed White people of almost 20%. In other words, for several decades at least, more unmixed White people will be leaving the labor market than are entering it.

The first-generation mixed minority–White group now makes up 9.4% of children under the age of 10 y (compared to 3.8% among the younger baby boomers). In evaluating its numerical size, some caution is required because of the convincing evidence that the reporting of mixed family backgrounds is selective in any single census dataset (31). This observation justifies using the ancestry variables to counteract this tendency. Despite selective reporting, the growth of the mixed group across birth cohorts is real because of the irrefutable evidence of increased mixing in families and among births (14).

The social positioning of the mixed group makes it a valuable window into assimilation processes that are more general but harder to capture with precision when examining unmixed minority categories. Moreover, that positioning for mixed young Americans does not end with their family ties to a White origin. By comparison with unmixed minority individuals, especially Black people and Hispanics, they are advantaged in multiple other ways, such as the higher average socioeconomic status of their parents and the greater likelihood that parents are married and reside in neighborhoods where many White people are present (14).

Birth certificate data, which show separately the ethnoracial categories of the parents, provide a more comprehensive picture of mixed backgrounds among children. Birth certificates have the advantage of data for many noncoresidential parents; however, when data are missing for fathers on birth certificates, we code the infants as unmixed on the assumption that the paternal link is likely broken. In 2019, mixed infants—here generalized to include all infants with parents from any two different major ethnoracial categories—constituted 14.7% of all births, or one of every seven. This group is dominated by infants who have one unmixed White parent (SI Appendix, Fig. S1) and one who is not. They make up about three-quarters of the whole, or 10.9%, approximately one of every nine babies.

The largest group by far among mixed infants is one that has received very little attention: one parent is unmixed White and one is Hispanic. This group is almost evenly divided between those with a Hispanic father and those with a Hispanic mother, and it accounts for more than half of all infants with mixed minority–White parentage. Other large first-generation mixed categories include infants of Black and White parentage (13.4% of all mixed infants), and those of Asian and White parentage (9.4%). The largest minority-only category of mixture is constituted by infants with one Black and one Hispanic parent (8.8%). Other minority-only categories are small.

Mixed minority–White infants form more substantial parts of each ethnoracial group (a mathematical necessity). Thus, more than one-fifth of infants with an Asian parent also have a White parent; the fraction is just slightly lower among those with a Hispanic parent. Likewise, almost one-fifth of infants with an unmixed White parent have one who is at least partly minority. The fraction of mixed minority–White infants is notably lower only among those with a Black parent (11.1%).

Along with the frequency of intermarriage, the proportion of infants with mixed parental backgrounds has been ratcheting upward over time. For example, in 2000, 8.2% of infants were from mixed parental backgrounds and 7.0% were mixed minority and White backgrounds. The percentage of mixed infants with minority-only backgrounds has been increasing rapidly, in tandem with the growing proportion of minorities among young adults. The frequency of babies with one minority parent and one White parent has also been growing steadily, and there is no reason to expect its growth to come to an end soon. The 2019 size of the minority–White mixed group among infants made it larger than the group of infants with an Asian parent (mixed or not) and about 60% of the group with a Black parent.

In sum, a demographic dynamic exists in the early 21st century that is likely to generate continuing mobility that brings Americans of minority descent (mixed with White descent or not) into contexts where few with non-White ancestry have been before. White people have been the dominant group in US society since the colonial period, and they have monopolized the upper strata, however defined, of American society ever since. Nevertheless, the size of the unmixed White group is declining with newer birth cohorts, setting up an ongoing situation in which there are not enough highly educated, unmixed White people to fill the desirable labor-market positions, whether opened by retirement or economic expansion.

Among those who are most likely to benefit from this opening is the growing number of mixed minority–White Americans, who are advantaged by their ties to White people and in other ways. This does not imply that they are the only persons of minority origins who are assimilating. But, of the groups definable in large-scale datasets, they are the most exemplary for assimilation processes.

Changes at the Top of the Socioeconomic Hierarchy

The changing ethnoracial demography and the increasing number of mixed minority–White youth are having sizable and growing impacts on the upper strata of the socioeconomic order, reflected in the rising diversity of college graduates and of the upper tiers of the US workforce. Demography is not the only force at work in these patterns: Legal and institutional changes since the civil rights period of the 1960s have opened avenues to higher socioeconomic attainment for non-White people through antidiscrimination laws and through affirmative action, which has been especially influential in higher education and in government employment (35). These changes have not brought about equality of life chances across the spectrum of ethnoracial origins, but they do mean at a minimum that non-White people are no longer barred from high-status spheres, as they were through most of American history.

As a prelude to the examination of the changing ethnoracial composition of the upper part of the workforce, we briefly describe its equivalent among college graduates, which is discussed elsewhere in detail (14). Baccalaureate attainment is a critical educational threshold because of its connection to high lifetime earnings (36). A comparison over time of baccalaureate rates and group sizes for individuals in their mid-20s reveals the dynamics that are changing the ethnoracial composition of the college graduates entering the labor market.

The educational life chances of unmixed White people remain superior to the aggregate of the rest of their birth cohorts. For example, among the young adults in the 2015 to 2019 American Community Survey, 34% of unmixed White men have a baccalaureate, compared to 20% of other men (SI Appendix, Table S4). Still, this disparity has declined relatively: In 2000, young White males, 28% of whom had a baccalaureate, were twice as likely to be college graduates compared to other young men.

There is great variation in the educational life chances of the others. As is well known, Asian Americans have high rates of college education, exceeding those of White people. However, others of unmixed minority origins have low rates. In general, the young adults from mixed minority–White backgrounds have rates in between those of unmixed White people and their respective minority groups. In the case of Asian–White people, this implies higher baccalaureate rates than those of White people. Of the other mixed groups, Hispanic–White people stand out, both because they are the largest mixed group and because their educational life chances are consistently much higher than those of unmixed Hispanics, though they are not quite at parity with those of unmixed White people.

These patterns, combined with the demographic shift in favor of young adults with some minority parentage, imply over time a declining share of unmixed White people in the pool of new college graduates. In 2000, they accounted for 71% of all graduates among those in their mid-20s; in 2015 to 2019, their percentage was down to 60% (37).

To be sure, the higher educational system is internally stratified and has an elite sector (e.g., the Ivy League) that has long been a training ground for the youth of the dominant group. However, White people no longer monopolize spots at elite schools. An examination of the composition of first-year classes over time shows the fraction of White people sinking at the top 100 colleges between 1980 and 2015, in many cases to 50% or below (38).

Employers who are seeking to hire new college graduates find a pool that is shifting in the direction of significantly greater ethnoracial diversity than would have been the case at the end of the 20th century. Moreover, relative to their unmixed minority peers, young Americans from mixed minority–White backgrounds are overrepresented in this diversity. These changes carry through to a decline in White dominance of the top tiers of the workforce. This dominance, which accommodated White–ethnic diversity after World War II (5, 39), has characterized US socioeconomic stratification since the nation’s founding.

To demarcate workforce tiers, we use the detailed occupations reported in census data, scoring them by the median annual incomes of their full-time workers.* Ranking occupations in this way and taking into account the numbers of their full-time incumbents, we identify those occupations that make up the top quartile of the full-time workforce. These occupations include all the major professions and other occupations that require advanced education, along with many technical jobs and supervisory and managerial ones, and even some highly skilled blue-collar positions: sales representatives and elevator installers are some nonprofessional examples.

Examining then the ethnoracial composition of the top-quartile occupations by birth cohort, we obtain a picture of sharply increasing diversity, in which the major groups of the post-1965 immigration play a powerful role. Table 1 shows the percentages in the top quartile of unmixed White, American Indian, Asian, Black, and Hispanic workers, as well as those from mixed minority–White backgrounds, by birth cohort in 2000 and 2015 to 2019. For the youngest category of the latter data, the minimum age for inclusion is 26 at the time individuals were surveyed.^†

Table 1.

Percentage of the incumbents of the top quartile of occupations, full-time workforce by ethnoracial category and birth cohort, 2000 and 2015 to 2019

Birth cohort:		1935 to 1944	1945 to 1954	1955 to 1964	1965 to 1974	1975 to 1984	1985 to 1993
Unmixed White	2000	85.8	82.9	78.3	73.1	—	—
	2015 to 2019	—	—	77.4	68.9	62.8	62.4
Unmixed Black	2000	4.4	5.5	6.5	6.9	—	—
	2015 to 2019	—	—	6.4	7.7	7.7	6.9
Unmixed Hispanic	2000	2.4	3.0	4.4	5.8	—	—
	2015 to 2019	—	—	4.7	7.2	8.7	9.1
Unmixed Asian	2000	3.5	3.9	5.2	7.7	—	—
	2015 to 2019	—	—	5.8	9.2	12.6	12.5
Unmixed Native American	2000	0.3	0.3	0.3	0.3	—	—
	2015 to 2019	—	—	0.3	0.3	0.3	0.3
Black–White	2000	0.1	0.1	0.2	0.2	—	—
	2015 to 2019	—	—	0.3	0.4	0.6	0.8
Hispanic–White	2000	0.4	0.5	0.7	0.9	—	—
	2015 to 2019	—	—	0.9	1.2	1.5	1.8
Asian–White	2000	0.2	0.2	0.4	0.7	—	—
	2015 to 2019	—	—	0.5	0.8	1.1	1.4
Native American–White	2000	1.4	1.7	1.8	1.8	—	—
	2015 to 2019	—	—	1.8	1.9	1.7	1.6
Other	2000	1.5	1.7	2.2	2.5	—	—
	2015 to 2019	—	—	2.0	2.4	3.0	3.3
Totals	2000	100%	100%	100%	100%	—	—
	2015 to 2019	—	—	100%	100%	100%	100%

Note: “—” indicates that the birth cohort is mostly or entirely out of working-age range (26–65 y). Calculations from the 2000 PUMS and 2015 to 2019 American Community Survey datasets (32–34).

For unmixed White people, the percentages in Table 1 going from the upper left, the oldest working-age cohort in 2000, to the far right in the next row, the youngest cohort in 2015 to 2019, map a shift from White dominance of the upper occupational strata to much greater diversity. More than 85% of the oldest top-quartile workers of 2000 were unmixed White people; and the slippage in the next cohort, the first baby-boom one, was modest. These cohorts, born in the late 1930s through the early 1950s, tell us about White dominance of the upper occupational tiers through much of the 20th century. But the fall-off in White dominance across birth cohorts is sharp. By the youngest two birth cohorts, the percentage of unmixed White people in the top tier of the occupational hierarchy has dropped to the low 60s.

The table also shows the ethnoracial composition of this increasing diversity. The biggest changes in composition are associated with workers of unmixed Asian and Hispanic backgrounds. Between the older baby-boom cohort (born 1945 to 1954) and the youngest workers (born 1985 or later), the percentages of unmixed Asian and of Hispanic workers each rose threefold; and in the latter cohort, the two groups combined account for more than one-fifth of the top-quartile workers. But they are distinguished by their nativity status. The Asian workers in the top quartile are dominated by immigrants, who still make up the majority of the Asian adult population, although the fraction of the United States–born is inevitably going up in tandem with the maturing of the second generation. The Hispanic workers are mainly United States–born (14).

The large minority whose presence is changing most slowly is African American. (Native American representation in the top quartile also has changed very little and remains well under 1%.) A comparison of the same two cohorts shows a small increase from 5.2 to 6.8%, although the share of unmixed Black workers reaches almost 8% in the 1965 to 1984 cohorts. What remains uncertain—and cannot be clarified with census data where the second generation, the children of immigrants, cannot be identified^‡—is the extent to which recent Black immigrants and their children may have better socioeconomic life chances than later-generation African Americans, who are mainly descended from enslaved ancestors; certainly, there is some evidence pointing in this direction, but it falls short of being definitive (10, 41–44). In any event, the disparity between Black and Hispanic representation, two groups often characterized as the most disadvantaged, seems largely a reflection of changing demography, as will become clearer below.

The mixed minority–White groups are also an important part of the story. Between the older baby boomers and the youngest workers, the mixed fraction of top-quartile workers more than doubled. The mixed group was not yet large among Americans born in the years of the youngest workers, 1985 to 1993, and we have shown earlier that its fraction has risen substantially in the most recent births. Based on this changing demography, it can be expected that mixed minority–White Americans will play a larger role in future diversity in the top tier.

A couple of potential misunderstandings should be avoided. The presence of different minority groups in the top quartile does not imply that they are distributed evenly across the occupations there, which manifest major differences in prestige and remuneration. For example, the percentages of unmixed Black people and of unmixed Hispanics among a top-quartile occupation’s incumbents are negatively correlated with its median annual earnings (−0.37, −0.36), while the percentage of unmixed Asians is positively correlated (+0.33). Of the unmixed groups, White people are the most evenly distributed: their percentage is minimally correlated (−0.08) with occupational status. The mixed groups are in two cases in between their minority peers and their White ones: that is, the percentage of Black–White workers in an occupation is negatively correlated with its remuneration (−0.23) but not as much as is true for unmixed Black workers, and an equivalent statement about a positive correlation (+0.16) holds for Asian–White workers. Hispanic–White workers are distributed, however, much like unmixed White workers (−0.05).

Moreover, unmixed White people are still overrepresented in the top quartile by comparison with their population shares of different birth cohorts. That is, their probability of holding a job there exceeds that of the unmixed minority groups, except for Asians. This is shown in Fig. 2, which is restricted to native-born workers in the 1955 to 1984 birth cohorts in 2015 to 2019. The restriction to the native born avoids—for Hispanics especially—a situation where the probabilities are confounded by the presence of many low-wage immigrants. The oldest and youngest cohorts in the data have been dropped because their probabilities are manifestly lower; in the case of younger workers, this occurs because many are just beginning their work lives.

Fig. 2.

Population-based probabilities (×100) of a top-quartile occupation by ethnoracial category and birth cohort, 2015 to 2019 Source: Calculations from the 2015 to 2019 American Community Survey (32, 34).

These probabilities reveal a strong stratification along ethnoracial lines, combined with modest improvements from older to younger cohorts. The groups with the highest overall probabilities are unmixed Asians and White people (the latter standing out in red in Fig. 2 for the purpose of comparison), along with two mixed groups: Asian– and Hispanic–White people. The lowest probabilities are associated with other unmixed minorities: American Indians, Black people, and Hispanics, along with mixed American Indian–White people. Black–White people and “others,” largely a mixed group including many individuals with complex or solely minority mixed ancestries, are in between.

Although the relative position of the groups is quite stable across the cohorts, there is an evident tendency for improvement in probabilities over time. This is obviously strongest for unmixed Asian workers, whose probability shoots up in almost linear fashion. In the case of unmixed White people, the probability improves from the oldest cohort in the figure to the youngest by about 20%. However, improvements tend to be larger for others: in the case of unmixed Black people, the comparable increase is 35%; for unmixed Hispanics, it is 25%. Large improvements are also the case for most of the mixed groups, such as American Indian–White (25%), Asian–White (33%), and Black–White people (24%). The one exception is Hispanic–Whites, whose probabilities already are very close to those of White people and experience a similar improvement of 19%.

The greater improvements in the probability of holding a top-quartile occupation for groups other than unmixed White people are part of the story for the increasing diversity at the top of the workforce. And of course, the high and increasing probabilities for workers of Asian descent play a role, too. But the improvements seem too modest, as does the size of the Asian American population, to explain fully the large changes in the composition of the top quartile evident across birth cohorts of workers.

A much bigger driver of change is demographic: that is, the changing ethnoracial composition of successive birth cohorts. The demographic changes are especially striking in a comparison of unmixed White people to the US-born contingents of the groups growing through immigration and to the mixed groups. Calculated from 2019 population data, the size of the native-born unmixed White group declined by 21% between the oldest and youngest cohorts in Fig. 2. In contrast, the size of the native-born groups of unmixed Asians and Hispanics increased by 82% and 69%, respectively. (The number of native-born unmixed Black people was stable.) Increases in the sizes of mixed groups are also generally large: Asian–White people grew in number by 150%, Black–White people by 112%, and Hispanic–White people by 69%. Only the American Indian–White group did not grow.

This section reveals apparently irrevocable shifts that are taking place in the upper spheres of the US socioeconomic order. To be sure, unmixed White people are still relatively advantaged in reaching these spheres. But, in large part due to a demographic shrinking of the White group and a shift toward more and more diverse cohorts of young Americans, there is growing diversity of ethnoracial background at the top of the occupational hierarchy, a fact that carries numerous ramifications for leadership in the US economy and other institutional domains. The increasing presence of Americans with minority family origins should be regarded as a form of social mobility, regardless of family socioeconomic status: For the individuals that are now present in social spaces associated with status and power and where few like them have appeared before.

This diversity at or near the top appears to favor those Americans with family origins in the immigration flows since the mid-20th century: unmixed Asians and Hispanics, most notably, and also those who come from mixed minority–White backgrounds. The presence of Black Americans is increasing but not to the same extent. There is no reason to think that this diversity will go backward, toward greater White dominance. If anything, the evidence about changes in the composition of the college educated augurs further increases in diversity at the top of the American workforce. The pool of young Americans obtaining college degrees is now very diverse.

Formation of Ethnoracially Mixed Families

A part of assimilation processes that is critical because of its impact both on individual lives and on the American population involves the formation of families that cross ethnoracial boundaries. This mixing can involve only minority origins, as exemplified by families formed by African American and Hispanic partners or by partners of different ethnic origins within a pan-ethnic category: for example, families combining Mexican and Salvadoran origins. But from the perspective of assimilation as a set of processes leading to mainstream incorporation, the most consequential family mixing unites White people with minorities, non-White and/or Hispanic people, since White people are, at this moment in US history, the dominant presence in the mainstream. This mixing is promoted by the processes of minority mobility sketched in previous sections, which bring White people and minorities together as equals in classrooms and workplaces.

Intermarriage has long been understood as a measure of change in the social distance between groups and hence of assimilation (15, 45, 46). Moreover, the best data we have about mixed families concerns intermarriage, although family formation does not require marriage and potentially could involve individuals who are not part of the same household but who maintain parental relations with the children they have had together. The intermarriage data show unambiguously that the rate of intermarriages has been rising steadily since the late 1960s, when the last of the state antimiscegenation laws were ruled unconstitutional; the national intermarriage rate was then about 3%. Today, almost one-fifth of new marriages cross major ethnoracial boundaries, and the great majority of these involve White and minority partners (13).

Our analysis of intermarriage recognizes the potential distinctiveness of the marriage patterns for individuals coming from mixed families. We have conducted separate marriage analyses for men and women, restricting them to native-born individuals under the age of 40 y (but including their spouses regardless of whether they meet these criteria). We consider only heterosexual marriages occurring in the last year in order to capture the most recent patterns and analyze the 2015 to 2019 American Community Survey data (though adding in the small number of same-sex marriages would not change the results).

Counting as intermarriages those where the ethnoracial heritages of the partners are not identical (hence, an intermarriage can bring together an African American partner with one who is from a mixed Black and White background), we find an intermarriage rate for both men and women of about 25 to 26%. For both sexes, almost 80% of the intermarriages involve a White partner, a fraction that is obviously higher than one would predict from the ethnoracial composition of the marriage-age population. It suggests (but does not prove) that marriage to a White American may be viewed as desirable by some coming from minority backgrounds. And many White people must want to marry them.

In Tables 2 and 3, we show the rates of marriage with unmixed White, mixed, and minority partners for individuals from different ethnoracial categories. Several observations seem pertinent. One is that the intermarriage rate by unmixed White people, at 15–16%, is higher than commonly shown in data (13). That is, distinguishing mixed minority–White partners shows that more White people are not strictly endogamous and are marrying into families with minority members.

Table 2.

Marriage patterns by ethnoracial categories of the partners, marriages during the past year, 2015 to 2019: Men

Husband’s ethno-racial category	Wife’s ethno-racial category (%)
	Unmixed White	Same or overlapping mixed*	Other mixed	Same unmixed minority	Other minority
Unmixed White	84.0	7.0	—	—	8.9
Unmixed Black	14.8	2.2	1.9	68.2	12.9
Unmixed Hispanic	23.2	2.3	2.7	65.1	6.6
Unmixed Asian	21.4	2.1	2.2	67.6	6.7
Unmixed Native American	37.7	6.9	1.3	40.5	13.7
Black–White	54.7	8.4	6.9	11.8	18.2
Hispanic–White	59.7	12.5	4.4	15.1	8.4
Asian–White	57.6	8.1	5.8	14.5	13.9
Native American–White	68.6	17.1	4.7	0.6	9.0
Other	41.6	—	9.5	20.8	28.2

Calculations from the 2015 to 2019 American Community Survey (32, 34). Focal individuals in each table are born in the United States and under 40 y of age; these restrictions do not apply to their partners.

*For unmixed categories, the column reports marriages to mixed minority–White individuals who share some ancestral background (e.g., Black people marrying Black–White partners); for mixed categories, the marriages are to partners with the same mixed background.

Table 3.

Marriage patterns by ethnoracial categories of the partners, marriages during the past year, 2015 to 2019: Women

Wife’s ethnoracial category	Husband’s ethnoracial category (%)
	White	Same or overlapping mixed*	Other mixed	Same unmixed minority	Other minority
Unmixed White	85.3	6.1	—	—	8.7
Unmixed Black	7.3	1.6	0.5	83.3	7.4
Unmixed Hispanic	20.8	2.1	2.1	66.6	8.4
Unmixed Asian	28.4	4.4	2.4	53.6	11.3
Unmixed Native American	40.4	2.7	1.9	38.2	16.8
Black–White	48.3	10.3	5.9	21.4	14.2
Hispanic–White	62.1	9.9	5.4	14.3	8.4
Asian–White	68.0	6.8	8.4	4.5	12.4
Native American–White	70.1	15.3	2.9	1.1	10.7
Other	46.0	—	7.8	18.5	27.8

Calculations from the 2015 to 2019 American Community Survey (32, 34). Focal individuals in each table are born in the United States and under 40 y of age; these restrictions do not apply to their partners.

*For unmixed categories, the column reports marriages to mixed minority–White individuals who share some ancestral background (e.g., Black people marrying Black–White partners); for mixed categories, the marriages are to partners with the same mixed background.

The second is the continuing overall low rate of marriage to White people by African Americans, which is differentiated by sex according to a historically well-established pattern (with Black men more likely to marry White partners than are Black women). Unmixed Black men in particular have a robust rate of outmarriage, about 30%; but their rate of marriage to unmixed White partners, at 15%, is much lower than one would expect from population proportions.

Third, other unmixed minorities, apart from the small Native American group, also have lower rates of marriage to White partners than the literature suggests. The drop of the rates shown in Tables 2 and 3 from those we would obtain using the conventionally defined census ethnoracial categories is particularly severe for Hispanic men and women (for men, from 31.0% to 23.2%; for women, from 29.2% to 20.8%). The cause is the inclusion of many mixed Hispanic–White individuals in the conventional Hispanic category; because they have high rates of marriage to White partners, when they are separated out, as they are by our mixed categories, the rates fall for those who remain in the unmixed Hispanic category.

More generally, this observation points to the overall importance of the marriage patterns of mixed individuals for intermarriage with White partners. All of the mixed categories that involve White ancestry have high rates of this form of intermarriage. Usually, the marriages to unmixed White partners far outnumber the marriages to members of the ancestral unmixed minority category. This is true, if to a modestly lesser degree, also for Black–White Americans, even though unmixed Black Americans have very low rates of marriage to White partners. The majority of Black–White men have taken unmixed White spouses, while the same is true for almost half of Black–White women. The figures are higher than in other recent data, suggesting growing mutual acceptability between Black–White and White Americans.

Conversely, marriage to persons of mixed minority–White background is important to the intermarriage rates of unmixed White people. The marriages of White men to partners of unmixed minority origins are almost the same in number as those to mixed minority–White individuals (7.0% to 8.9%), but in the marriage-age population, minority-only persons are far more numerous than mixed ones. A similar pattern appears for White women.

Minority intermarriage to other than unmixed White individuals does not exhibit very salient features. In general, the endogamy rates of unmixed minority categories are lower than that of White individuals, a difference driven in part by relative population size, although that for Black women, at 83.3%, is considerably higher than the rest. The lowest endogamy rate is for Native American women, at 38.2%; that for Native American men is only slightly higher. The endogamy rates for the mixed minority–White categories are generally quite low: the highest is 17.2% for Native American–White men.

In general, the rates of marriage between individuals of different minority origins are also low. This is partly a function also of population sizes, especially in light of the robust endogamy rates of the unmixed categories. Of the intermarriages rates to other persons of minority background, the strongest pattern is one of marriage to unmixed Hispanics, reflecting in part the relative size of the Hispanic group among non-White individuals. For example, 10.7% of Asian–White men have taken Hispanic partners, as have 10.3% of Black–White men.

In conclusion, marriage patterns confirm that mixing across the major categorical boundaries of race and ethnicity, especially between White people and others, remains robust. The children of mixed families formed in the past are critical to mixing as we observe it in census data. That is, individuals of mixed minority–White backgrounds have high rates of marriage to unmixed White individuals, and in tabulations that do not distinguish sufficiently between mixed and unmixed backgrounds, their presence in single-race or Hispanic categories inflates somewhat the apparent intermarriage rates with White partners. The high rates for mixed minority–White Americans are indicative of their integration into social milieus—such as job sites, school friendship circles—that, though they may be diverse in many cases, contain many White people, for these are the environments in which many Americans find mates.

Since unmixed White Americans are still the majority of the population, their marriage patterns are especially important in determining the overall frequency of intermarriage. The evidence is compelling that their rates of endogamy are declining gradually. In our data, where individuals of mixed White and minority family backgrounds have been separated from those who are unmixed White individuals, the intermarriage rates of the latter have reached to about 15–16%, clearly above what they have been in the past but far below what they would be in a world where race and Hispanic ethnicity did not represent salient social distinctions. Nevertheless, as demographic change leads to youthful cohorts that are increasingly diverse, this rate will almost certainly go up because White people will encounter more peers who are not White in background or not White alone.

Conclusion

In the social sciences, ideas associated with racial stratification are casting a long shadow over discussions in the United States about the social locations of immigrants who have come since the 1965 immigration-law revision and their descendants. This is the case because the great majority of immigrants have come from countries outside of Europe, with the largest recent numbers from countries in Asia and Latin America. Guided by the precepts of critical race theory, many social scientists envision American society as fundamentally organized around a rigid racial hierarchy, which is maintained by major social institutions and the actions of many individuals for the benefit of the dominant group, White Americans. Since the immigrants are mostly regarded as not White according to the ethnoracial classifications of the US census, it follows that they and their descendants should fall under the scope of race-based theories.

However, the evidence presented here does not fit with this expectation and supports the notion that many immigrants and their descendants are pursuing mobility strategies that are linked to integration into mainstream settings, where many White people are also present. To be sure, that most of the post-1965 immigration came after legal and institutional changes initiated by the Civil Rights Era has altered fundamentally the contexts in which minority individuals devise mobility strategies. The conclusion that the evidence about mobility and social integration is inconsistent with race theories is not limited to the descendants of immigrants; it includes other minorities, such as African Americans, too, who are mainly descended from ancestors brought in slavery. Minority mobility strategies, when successful, often lead to forms of assimilation, whether this is intended or not. Although the empirical focus of this article has been most concentrated on Americans from mixed minority–White family backgrounds, assimilation processes are not restricted to them, since social mobility and social integration are evidently much broader. The analysis of the composition of the top of the workforce, for instance, demonstrates the growing presence of unmixed minorities and implies increasing equal-status contacts between the White group already established there and the mobile non-White and Hispanic groups.

Yet, the limitations to these processes must also be recognized. Assimilation is most pertinent to individuals from immigration backgrounds and appears not as extensive for other minorities, African Americans, above all. The greater relevance of assimilation for immigrant-origin Americans is likely the product of several distinct forces: the selectivity of many immigrant streams, the greater optimism of immigrants and their children, and their greater acceptance by members of the dominant group, White Americans. Even though the Black–White boundary has proven stubbornly durable, there is suggestive evidence there, too, that it may be more permeable for recent Black immigrants and their children. Yet, we cannot understand the positions of all immigrants and their descendants only with assimilation ideas. To state the most obvious: immigrants who are unauthorized and their children, even US-citizen children, are disadvantaged in ways that are not illuminated by assimilation (47). Other immigration-origin Americans are disadvantaged by their racial distinctiveness, by Black or Indigenous skin tones and features (48).

Assimilation ideas then do not replace race-based theories but coexist with them. Both are essential for any understanding of contemporary sociological realities. Assimilation holds a special importance for the ways we should think about the changes taking place in an ethnoracially stratified but increasingly diverse society. Americans’ conception of the future has been very much influenced by the majority–minority forecast of the future arising from raced-based thinking and census data, according to which US society is inevitably heading toward a future in which White people will be a numerical minority and ethnoracial minorities collectively constitute the majority (4). However, the evidence presented here and elsewhere for mixed minority–White Americans shows that they do not fit this binary conception and undermines this forecast of a society split in two along a single ethnoracial fault line.

A large question still hovers over the long-run outcome of these processes: Do they imply a “Whitening” of assimilating minorities, and thus an augmented White population, or will the mainstream part of society become less visibly White and more diverse (49)? There cannot be a conclusive answer to such a question, and undoubtedly the future will represent a blend of Whitening and growing diversity. But history offers a significant clue: When the so-called White ethnics, mainly Catholic and Jewish, assimilated en masse in the middle of the last century, they did not simply merge into the dominant group, Anglo-American Protestants. They retained their religious distinctiveness and asserted their own ethnic identities, although both were modified in ways to reduce conflicts in everyday social relations that frequently crossed ethnic and religious lines. Today, also, we do not see the extinction of distinct identities among assimilating minorities; even mixed, partly White Americans frequently reference their minority origins (31, 50); and the mainstream culture appears open to influences from contemporary minorities, especially of immigrant origin (51). If the future follows the past model, as seems plausible, then assimilation will gradually tilt the mainstream away from exclusive Whiteness. Obviously, this will not mean by itself an end to problems of racial inequality, but it still seems a welcome development.

Data Availability

All study data are included in the main text and SI Appendix.