California's Multiracial Population

Public Policy Institute of California California Counts POPULATION TRENDS AND PROFILES

Hans P. Johnson, editor Volume 6 Number 1 • August 2004

California’s Multiracial Population By Laura E. Hill, Hans P. Johnson, and Sonya M. Tafoya

Before Census 2000, Americans were asked to choose just one race when identifying themselves and their children. With the advent of ummary the option to choose one or more races in Census 2000, there was a great deal of uncertainty about just how many Americans consider themselves to be multiracial. As with other issues related to racial and ethnic diversity, California is leading the nation—5 percent of the state’s population is identified as being of more than one race, about twice the rate as in the rest of the nation. In this issue of California Counts, we explore this newly identified population. We find that California’s multiracial population is hard to characterize with any basic summary statis- tics. Overall, people who identify themselves as multiracial are younger, less educated, slightly more likely to be foreign-born, and more likely to be living in poverty than single-race Cali- fornians. However, multiracial Californians are of many racial combinations, with very dif- ferent characteristics according to the particular combination. For example, the median age of individuals identified as both black and white is only 12 years, compared to 36 years for American Indian and white Californians. The poverty rates for individuals identified as Asian and white is less than half that of Hispanics who identify as both white and some other race. For the most part, biracial Asian and whites, American Indian and whites, and black and whites have socioeconomic characteristics intermediate to those of their monoracial counter- parts. However, both black and whites and Asian and whites are significantly younger than their monoracial counterparts, suggesting that the characteristics of the multiracial population could change as more and more children are born to parents of different races and potentially retain multiracial identity as they grow into adulthood and have their own children. In the near term, the presence of this new multiracial option presents some challenges for the collection and analysis of demographic data at the state and national levels. We already see Public Policy Institute of California

California Counts California’s Multiracial Population

As with other issues evidence that demographic rates calculated using different data sources can related to racial and lead to implausible results for multiracial populations. Ultimately, the size ethnic diversity, Cali- and significance of the multiracial population of California will depend at fornia is leading the least partly on future preferences with respect to identity. The ability to choose more than one race on state forms and future censuses along with nation—5 percent of increasing rates of intermarriage could lead more Californians to choose a the state’s population multiracial identity. As the multiracial population grows, it has the power is identified as being to challenge and even transform our understanding of race in California. of more than one race, about twice the rate Laura E. Hill and Hans P. Johnson are research fellows at the Public Policy Institute of as in the rest of the California. Sonya M. Tafoya is a research associate at the Pew Hispanic Center. The authors nation. acknowledge the helpful comments and thoughtful reviews of Gary Bjork, Kyra Greene, Aaron Gullickson, Mary Heim, and Deborah Reed. Views expressed are those of the authors and do not necessarily reflect the views of PPIC or the Pew Hispanic Center.

2 Public Policy Institute of California

California Counts California’s Multiracial Population

Introduction identity in California and the rest Is there such a thing of the nation, identifying which as multiracial identity? ive percent of California’s resi- biracial categories are the most Will these new race Fdents are multiracial, which is common, comparing the demo- double the percentage recorded at graphic characteristics of biracial options change the way the national level (2.4%), according Californians to those who choose we think about racial to Census 2000. These 1.7 million only one race, and describing the and ethnic identity in multiracial Californians represent socioeconomic characteristics of the future? 25 percent of the nation’s multi- biracial Californians, including racial individuals. Thus, to under- their political attitudes. We then stand who multiracial Americans conclude with a discussion of the are and what it means to identify as future of race in California.1 multiracial, it makes sense to start with California. Because Americans The Census Question and Californians have only recently During the debate about whether “mark one or more races to indi- had the opportunity to officially to allow Americans to identify cate what this person considers state a multiracial identity (begin- themselves as multiracial, there himself/herself to be.” Hispanics ning with Census 2000), little is were many different proposals for may identify with any racial group known about those who choose a how to ask the question. Would or combination of groups.2 biracial or multiracial identity. people be asked to indicate which As shown in Figure 1, there An individual may be identified races they identified with or only are 15 race boxes on the census as being of more than one race for to indicate that they were “more form. Asians, for example, are sub- at least two reasons. Some individ- than one race” or “multiracial”? divided into Asian Indians, Chi- uals are identified as multiracial After much debate, it was decided nese, Filipinos, Japanese, Koreans, because they have parents or ances- that the same race question would Vietnamese, and Other Asians. tors of different races. Other indi- be used as in past censuses, but Although there is no indication viduals are identified as multiracial that in Census 2000, people could on the census form that someone because they do not believe that the choose as many race boxes as they who selects Filipino and Japanese standard single-race categories ade- wished, as shown in Figure 1. will not be tabulated as multiracial, quately describe their racial identity. As in the past, the Office of such respondents are tabulated as Although we cannot distinguish Management and Budget (OMB) monoracial Asian for official tabu- between these two reasons in Cen- and consequently the Census lations. Or, for example, if some- sus 2000 data, we can ask who Bureau view Hispanic origin as an one checked the boxes indicating multiracial Californians are and how ethnic identity rather than a race. Native Hawaiian, Guamanian or they differ from those who have The result of this distinction is that Chamorro, Samoan, and Other only chosen one race. Is there such census data collection methods Pacific Islander, he or she would a thing as multiracial identity? Will consist of two related questions: not be counted as belonging to these new race options change the a Hispanic ethnicity question and a four race groups, but rather would way we think about racial and eth- race question. Question 5 records be tallied as monoracial Native nic identity in the future? an individual’s Hispanic ethnicity, Hawaiian or Other Pacific Islander In this issue of California if any, and Question 6 records races, (NHOPI). Although Figure 1 out- Counts, we examine multiracial instructing census respondents to lines the broad racial categories,

3 Public Policy Institute of California

California Counts California’s Multiracial Population

it does not specify who will be tal- Figure 1. Reproduction of Questions on Race and lied as multiracial or how write-in Hispanic Origin, Census 2000 responses are treated. If a respon- dent writes in an ethnic identity, ➔ he or she is often recoded into one NOTE: Please answer BOTH Questions 5 and 6. of the five standard race categories 5. Is this person Spanish/Hispanic/Latino? Mark ✗ the “No” box if not Spanish/Hispanic/Latino. (white, black, Asian, NHOPI, or No, not Spanish/Hispanic/Latino Yes, Puerto Rican some other race [SOR]). For exam- Yes, Mexican, Mexican Am., Chicano Yes, Cuban ple, those who write-in “Mexican” Yes, other Spanish/Hispanic/Latino — Print group. are recoded into SOR.3

Multiracial Population 6. What is this person’s race? Mark ✗ one or more races to indicate what this person considers himself/herself to be. Sizes in California and the White United States Black, African Am., or Negro Numerically, California is clearly American Indian or Alaska Native — Print name of enrolled or principal tribe. home to the greatest number of multiracial Americans of any state and has a disproportionately large Asian Indian Japanese Native Hawaiian multiracial population given its Chinese Korean Guamanian or Chamorro size (Table 1). Filipino Vietnamese Samoan Other Asian — Print race. Other Pacific Islander — Print race. Other big states with large immigrant populations and diverse populations also have large num- Some other race — Print race. bers of multiracial residents (e.g., New York, Texas, and Florida). However, Hawaii, which has a very Source: U.S. Census Bureau, Census 2000 questionnaire. small total population (smaller even than California’s multiracial population with only 1.2 million residents), has the fifth largest number of multiracial residents by state. Indeed, as a percentage of Table 1. Top Five States Ranked by Number and the population, Hawaii ranks the Percentage of Multiracial Residents, Census 2000 highest, with more than one in Rank Number Multiracial (millions) Percent Multiracial five residents identified as multira- cial. California ranks third among 1 California 1.6 Hawaii 21.4 states by the percentage of multi- 2 New York 0.6 Alaska 5.5 racial residents. 3 Texas 0.5 California 4.8 What races do multiracial Cal- 4 Florida 0.4 Oklahoma 4.5 ifornians choose and in what pro- 5 Hawaii 0.3 Nevada 3.8 portion? How do the chosen races Source: Authors’ tabulations using Census 2000 1% Public Use Microsample (PUMS). compare to those chosen by multi-

4 Public Policy Institute of California

California Counts California’s Multiracial Population

racials nationwide? Where is the multiracial population located Text Box 1. Key to Abbreviations and Usage throughout the state and how does this vary by multiracial group? Do California’s multiracial individuals Race categories and abbreviations for single-race groups differ in any substantive way from White = white their monoracial counterparts in Black = black or African American their demographic or socioeco- Asian = Asian nomic characteristics? The follow- AIAN = American Indian and Alaska Native ing sections of this report address NHOPI = Native Hawaiian and Other Pacific Islander each of these questions in turn. SOR = some other race Multiracial = person belonging to more than one of the above six groups. Choose Which Biracial = person belonging to two of the above six groups. Two? California’s Multiracial and Conventions for population of Hispanic origin The terms “Hispanic” and “Latino” are used interchangeably Ethnic Composition throughout the text. Both terms refer to the same demographic group. he majority of Californians who checked more than one T Conventions for referring to the population of multiracial race on the Census 2000 form Americans chose just two races. Indeed, less A “+” sign is used throughout the text of this report when referring than half a percent of Californians to individuals identified as more than one race. For example, a per- indicated being of three or more son who marked the checkbox for white and for Asian is referred races (0.3%). Five biracial combi- to as Asian+white. nations account for more than 71 percent of all the state’s multi- Because Hispanic ethnicity is a separate census question, a person racial population (see Text Box 1 who checked any Hispanic ethnicity in Question 5 and two or for a guide to race conventions more races in Question 6 would be a Hispanic multiracial individ- used in this report). ual. For example, a person who is Hispanic and checked both white The most common multiracial and SOR would be referred to as Hispanic SOR+white. group is Hispanics who are identi- fied as both white and SOR (see Conventions for referring to the race(s) of married couples Table 2). A “/” is used throughout the text of this report when referring to California’s next most common the racial identification of married couples. For example, a couple multiracial group is non-Hispanic in which one partner is black and the other partner is Asian is SOR+white. This differs from referred to as a black/Asian couple. If one partner is multiracial, the national ranking where non- black+Asian, and the other partner is white, the couple would be Hispanic AIAN+white is the sec- designated black+Asian/white. ond most common multiracial category. In California, the non-

5 Public Policy Institute of California

California Counts California’s Multiracial Population

Hispanic Asian+white group also Table 2. Most Frequent Biracial Combinations, outnumbers the non-Hispanic California, Census 2000 AIAN+white group. As in the country as a whole, non-Hispanic California Percent of black+whites are the fifth most California Population California National Rank Biracial Group Size Population Rank common biracial group. In the remainder of this report, all races 1 Hispanic, SOR+white 483,982 1.4 1 without “Hispanic” preceding 2 Non-Hispanic, SOR+white 238,473 0.7 4 them are non-Hispanic. 3 Non-Hispanic, Asian+white 232,106 0.7 3 Overall, monoracial groups are 4 Non-Hispanic, AIAN+white 153,982 0.5 2 the most common in California, 5 Non-Hispanic, black+white 91,888 0.3 5 with whites, Hispanic SORs, His- panic whites, Asians, and blacks Source: Authors’ tabulations using Census 2000 5% PUMS. composing the top five racial groups and over 93 percent of California’s population (Table 3). Table 3. Top 15 Racial and Ethnic Categories in However, three biracial California, Census 2000 groups—Hispanic SOR+white, Percent of SOR+white, and Asian+white— Rank Category Number California Population rank sixth, seventh, and eighth, 1 White 15,787,039 46.6 respectively, and each above single- 2 Hispanic SOR 5,640,156 16.6 race AIANs. The five most com- 3 Hispanic white 4,339,673 12.8 mon biracial pairings each vary 4 Asian 3,661,295 10.8 tremendously in their size relative 5 Black 2,161,542 6.4 to their monoracial components. 6 Hispanic SOR+white 483,982 1.4 For example, the number of 7 SOR+white 238,473 0.7 black+whites is less than 5 percent 8 Asian+white 232,106 0.7 of the size of the black population 9 AIAN 188,221 0.6 (and a much smaller share of the 10 AIAN+white 153,982 0.5 white population). On the other 11 Hispanic AIAN 129,425 0.4 hand, SOR+whites number more 12 NHOPI 94,925 0.3 than three and a half times the 13 Black+white 91,888 0.3 size of the monoracial population 14 Hispanic black 70,231 0.2 of SORs. Asian+white Californians 15 SOR 67,280 0.2 are approximately 6 percent of the size of the monoracial Asian pop- Other 544,448 1.6 ulation. The size of the Hispanic SOR+white population is approx- Total 33,884,660 100.0 imately 11 percent of each the Hispanic SOR and the Hispanic Source: Authors’ tabulations using Census 2000 5% PUMS. Note: Percentages may not sum to 100 because of rounding; categories are exclusive by race and white population. Finally, the ethnicity (e.g., “white” refers to non-Hispanic whites). AIAN+white population is over

6 Public Policy Institute of California

California Counts California’s Multiracial Population

80 percent of the size of the monoracial AIAN population. Text Box 2. Consistency in Racial Reporting The SOR race category merits further discussion here. The inclu- sion of the SOR category among Racial identification is not immutable. Even for the same individ- the standard single-race categories ual, it can change with context, over time, and even according to allows the Census Bureau to enu- the format and order of questions on surveys. Among multiracial merate those who do not identify individuals, consistent reporting of multiracial identity is quite with the racial categories that the low. For example, among non-Latinos who reported two or more census employs. The majority races in Census 2000, only 40 percent responded by offering the (97%) of those who select SOR same response in a follow-up survey conducted by the Census alone were Hispanic (Grieco and Bureau. Of the 60 percent selecting only one race in the follow-up, Cassidy, 2001, Tafoya, 2003). most selected white. Seventy-five percent of non-Latino biracial Among all who identify as SOR+ Americans identifed as at least partly white in Census 2000. Con- white, two-thirds identified as eth- sistency in racial reporting is lower among multiracial Latinos: nically Hispanic. The non-Hispanic Only 22 percent of Latinos who identified as multiracial in Census third of this group reported an 2000 also reported two or more races in the Census Bureau’s array of different ancestries includ- follow-up. More than 29 percent selected SOR alone, and even ing Armenian, Iranian, Italian, more—43 percent—selected white alone. The two multiracial Portuguese, Arab, Irish, German, groups that were most consistent in their reporting of race across and Egyptian, on a separate census the two data collection efforts were Asian+whites and black+whites (Bennett, del Pinal, and Cresce, 2003). question that elicits ancestry. The In the California Health Interview Survey (CHIS)—a survey identity of these respondents may conducted by telephone—multiracial respondents were asked reflect mixed parentage, an under- which race they most identified with. When asked to choose one standing of race at odds with cen- race, black+whites were more likely than any other biracial group sus categories, or an ancestry that studied to choose a nonwhite response. Among black+whites, is nominally white but an experi- almost half chose black or African American, only 18 percent chose ence in the society at large as only white, and one-third chose something else (including multiracial, marginally white. For this latter but this response cannot be separated from the others). Among group, it is difficult to argue that Asian+whites, 42 percent selected white, 29 percent chose Asian, they share a racial identity given and 29 percent chose something else. Among AIAN+whites, the the diversity of ancestries that vast majority identified with white (75%). Many fewer AIAN+ they acknowledge, but they clearly whites than Asian+whites or black+whites—9 percent—persisted in share a preference for avoiding the choosing something else (including multiracial). Because “other” standard monoracial categories, was not read as one of the racial responses in the CHIS survey, very including monoracial SOR. few Latinos identified as SOR+white in the CHIS. There is some question about how consistently these individuals identify as multiracial. In Text Box 2, we discuss the degree of variation in racial identification for multi- racial Americans.

7 Public Policy Institute of California

California Counts California’s Multiracial Population

Where Are Those Who Choose More Map 1. Percentage Multiracial, by County, Census 2000 Than One Race? ithin California, multiracial Wresidents are distributed somewhat unevenly, with higher proportions in some of the state’s rural and smaller metropolitan counties, in addition to some of the state’s urban counties (Map 1). The majority of the state’s most multiracial counties (those with more than 5 percent multiracial residents) are in the northern San Joaquin Valley, the Sacramento region, and the Bay Area. Sacra- mento County is the largest urban county that ranks high in the share of its residents that are mul- tiracial (5.8%). Earlier research on <3.0% births to mixed-race couples in 3.0–3.9% California also found that these 4.0–4.9% 5.0%+ births were more likely in Sacra- mento County than in other regions of the state (Tafoya, 2000). Solano County has the highest Source: Authors’ tabulations using Census 2000 SF1. percentage multiracial residents, with 6.4 percent. What is especially notable about California’s multiracial population population shares of 7 percent or are SOR+white, with ancestry is how few of the state’s 58 coun- higher (Table 4). data indicating many of Armenian ties have less than 3 percent of More than 10 percent of descent. Newport Beach, in South- their population that is multiracial Southern California’s Glendale ern California, has the lowest per- (recall that the national average population is multiracial, as is over centage of multiracial residents was 2.4%). Indeed, only Mono 7 percent of the population in a (1.7%). County has a lower proportion of number of cities in the wider San Because Hispanic SOR+whites its residents that are multiracial Francisco Bay Area (Hayward, are the most common multiracial than the national average (2.2%). Fairfield, Pittsburg, South San group statewide, they also tend to The six most multiracial cities in Francisco, and Antioch). In Glen- dominate the multiracial popula- the state each have multiracial dale, most multiracial residents tion in any given locale. When

8 Public Policy Institute of California

California Counts California’s Multiracial Population

The six most multiracial Table 4. California’s Most and Least Multiracial Cities, Census 2000 cities in the state Most Multiracial Least Multiracial each have multiracial

Percent Percent population shares of 7 percent or higher. Glendale 10.1 Newport Beach 1.7 Hayward 7.5 Thousand Oaks 2.8 Fairfield 7.4 Encinitas 2.9 Pittsburg 7.2 Carlsbad 3.0 South San Francisco 7.0 Yorba Linda 3.1 Antioch 7.0 Cupertino 3.1 we examine California’s ten largest Source: Authors’ tabulations using Census 2000 5% PUMS. cities (Table 5), we find that His- Note: The table includes cities with at least 50,000 residents. panic SOR+whites are the most common multiracial group in nine of them. San Francisco, California’s tenth largest city, is the one exception, Table 5. Multiracial Populations in California’s Ten where Asian+whites are the most Largest Cities, Census 2000 common multiracial group. Los City Multiracial Angeles, the largest city in the state, Population Share of City (%) Largest Multiracial Group has the greatest number of multi- Sacramento 407,018 6.4 Hispanic SOR+white racial individuals of any city state- Long Beach 461,522 5.3 Hispanic SOR+white wide, and this is true for each of the Los Angeles 3,694,820 5.2 Hispanic SOR+white five most common biracial groups. Fresno 427,652 5.2 Hispanic SOR+white San Jose 894,943 5.0 Hispanic SOR+white Anaheim 328,014 5.0 Hispanic SOR+white Demographic Oakland 399,484 5.0 Hispanic SOR+white San Diego 1,223,400 4.8 Hispanic SOR+white and Socioeconomic Santa Ana 337,977 4.6 Hispanic SOR+white Characteristics San Francisco 776,733 4.3 Asian+white of Multiracial Source: Authors’ tabulations using Census 2000 5% PUMS. Californians n the remainder of this issue of ICalifornia Counts, we focus on the five most common multiracial responses among Californians (those listed in Table 2) and com- pare their characteristics to those of their monoracial counterparts.

9 Public Policy Institute of California

California Counts California’s Multiracial Population

Here, we examine nativity and age more likely than SORs to be nians, who have a median age of distributions, then turn to socio- foreign-born, and this group is 12, compared with over age 30 for economic characteristics. largely responsible for the higher either monoracial blacks or mono- Multiracial Californians are share of foreign-born overall among racial whites, suggesting that the slightly more likely than monora- the multiracial population than multiracial identities among this cial Californians to be foreign-born among the monoracial population. group are being taken up primar- (28% versus 26%). However, this Multiracial individuals, as a ily by the very young, as reported varies from one biracial group to group, are younger than their mono- by their parents.4 the next. Black+whites are less racial counterparts. Of the 6.8 Asian+white biracial individu- likely to be foreign-born than either million Americans who reported als are also quite young relative to whites or blacks, as are AIAN+ more than one race, 42 percent both white and Asian monoracial whites relative to AIANs and were under age 18, compared to groups, with a median age of 18. whites, and Hispanic SOR+whites only 25 percent for monoracial Both AIAN+white and SOR+ relative to Hispanic whites and Americans (Jones and Smith white median ages fall between Hispanic SORs (Figure 2). 2001). In California, we find that those of each monoracial group. Asian+whites are much less multiracial people have a median Hispanics as a group are relatively likely than Asians, but more than age ten years lower than that of young, and because Hispanic SOR+ twice as likely as whites, to be the monoracial population (24 whites constitute the largest multi- foreign-born. SOR+whites are five versus 34) (Figure 3). The most racial group (nearly 0.5 million), and a half times more likely than striking age difference is the rela- they contribute strongly to the whites and one and a half times tive youth of black+white Califor- overall age structure of the broader multiracial population. Age dif- ferences by gender within each biracial category are few and small, Figure 2. Percentage Foreign-Born, Census 2000 but, on average, females appear to be the same age or a year or two 80 older than their male multiracial

70 counterparts (not shown). These differences are examined more 60 closely in age pyramids for the two 50 biracial groups with the youngest 40 median age (black+white and

Percentage 30 Asian+white) as well as their cor- 20 responding monoracial groups 10 (Figure 4).

0 The population pyramid for black+whites is striking in its SOR White White Asian White AIAN White Black youth relative to either whites or MultiracialSingle race SOR+white Asian+white AIAN+white Black+white SOR, Hispanic blacks. Indeed, it resembles the White, Hispanic population pyramids found in SOR+white, Hispanic developing countries, where birth Source: Authors’ tabulations using Census 2000 5% PUMS. rates and mortality rates are high.

10 Public Policy Institute of California

California Counts California’s Multiracial Population

Because interracial Figure 3. Median Age, Census 2000 couples are more common recently than 45 40 in the past, the biracial 35 offspring of these cou- 30 ples will be relatively 25 young. 20

Percentage 15 10 5 black+white respondents chose black than white, whereas among 0 Asian+whites, many more chose SOR White White Asian White AIAN White Black white than Asian (authors’ tabu- MultiracialSingle race SOR+white Asian+white AIAN+white Black+white SOR, Hispanic lations from the CHIS, 2002). White, Hispanic Finally, many parents of older SOR+white, Hispanic children may not avail themselves Source: Authors’ tabulations using Census 2000 5% PUMS. of the opportunity to check more than one box on the Census 2000 form because they have had to The population pyramid for als and their parents might feel choose just one race box on so Asian+whites is similar, although less constrained by traditional many other forms in the years the decline in population size at racial categories that have histori- since their children were born. older ages is not quite as dramatic cally forced people to choose a Overall, the multiracial adult as that for black+whites. single racial identity. However, population (age 25 and older) is Other forces than fertility and these children and their parents slightly less educated than the mortality are shaping the extreme may realize later from experiences monoracial population (Figure 5). population pyramids observed for in school and in the world at large However, for most groups of bira- Asian+whites and black+whites. that they are perceived as monora- cial Californian adults, the level of First, to the extent that interracial cial. This may cause these children educational attainment is between couples are more common recently or their parents to switch to a that of their component monora- than in the past, the biracial off- monoracial identity as the children cial groups. spring of these couples will be rel- age. Indeed, the steepness of the For example, black+whites atively young. Indeed, the share sides of the black+white popula- tend to be better educated than of all births that are to mixed-race tion pyramid relative to the Asian+ blacks but less educated than couples has increased approxi- white pyramid suggests that more whites. The percentage of these mately 40 percent in the last 20 black+white children over the age populations with less than a high years in California, from 5.3 per- of ten are identifying as monora- school education is 10 percent for cent in 1982 to 7.4 percent in cial. When multiracial Californians whites, 13 percent for black+whites, 2001 (Tafoya, Johnson, and Hill, were asked to select their preferred and 19 percent for blacks. Like- 2004). Second, younger individu- monoracial category, many more wise, 34 percent of whites have a

11 Public Policy Institute of California

California Counts California’s Multiracial Population

Figure 4. Population Age Pyramids, Census 2000

Black Asian 90+ 90+ 85 85 80 80 75 75 70 70 Male Female 65 Male Female 65 60 60 55 55 50 50 45 45 Age Age 40 40 35 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0 150 100 50 0 50 100 150 200 150 100 50 0 50 100 150 200 Population size (thousands) Population size (thousands)

Black+white Asian+white 90+ 90+ 85 85 80 80 75 75 70 70 65 65 60 60 55 Male Female 55 Male Female 50 50 45 45 Age Age 40 40 35 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0 150 100 50 0 50 100 150 25 20 15 10 5 0 5 10 15 20 25 Population size (thousands) Population size (thousands)

White 90+ 85 80 75 70 Male Female 65 60 55 50 45 Age 40 35 30 25 20 15 10 5 0 800 600 400 200 0 200 400 600 800 Population size (thousands) Source: Authors’ tabulations using Census 2000 5% PUMS.

12 Public Policy Institute of California

California Counts California’s Multiracial Population

Figure 5. Educational Attainment, by Racial Group, We find that multi- Census 2000 racial people have a median age 10 years Less than high school High school diploma Some college Bachelor’s 100 lower than that of the 90 monoracial population 80 (24 versus 34). 70 60 50 40 Percentage 30 These high poverty rates can be 20 at least partly attributed to a gen- 10 erally young age structure and a 0 relatively high share of foreign-born

SOR individuals in the state’s multiracial White White Asian White AIAN White Black population. In California, poverty MultiracialSingle race SOR+white Asian+white AIAN+white Black+white SOR, Hispanic rates tend to be higher for house- White, Hispanic holds with children than for house- Source: Authors’ tabulations using Census 2000 5% PUMS. SOR+white, Hispanic Note: Ages 25+. holds without children (Reed, 2004). Immigrants tend to have lower earnings and thus higher bachelor’s degree; among black+ Both Hispanic SOR+whites poverty rates than U.S.-born resi- whites, 30 percent are in this cate- and non-Hispanic SOR+whites dents of the state (Reed, 2004). gory, and among blacks alone, only are less educated than either of However, poverty rates vary 17 percent have attained at least a their monoracial counterparts. widely between the largest multi- bachelor’s degree. The same pattern Differences among Hispanic SOR+ racial groups. For example, poverty prevails when comparing white, whites, Hispanic whites, and His- rates for Asian+whites are less AIAN+white, and AIAN groups. panic SORs are slight, however, than half those of SOR+white His- Whites show the highest educa- and those among SOR+whites, panics (8% versus 22%) and are tional attainment, AIAN+white SORs, and whites are substantial. substantially lower than the state biracials are intermediate, and the Fifty-four percent of SOR+whites poverty rate. With one exception, AIAN group has the lowest edu- attended college versus 69 percent poverty rates for biracial groups cational attainment. In the com- of whites and 63 percent of SORs. are intermediate between those parison of whites to Asian+whites To measure economic well- of their monoracial components. and Asians, the biracial group is being, we examined poverty rates For example, black+white poverty again intermediate although very for all ages of multiracial Califor- rates are halfway between the rela- similar to whites; however, the nians. As with educational attain- tively low rates of whites and the monoracial Asian group has the ment, we found that Californians relatively high rates of blacks. The highest level of educational attain- identified as multiracial are slightly one exception is among the SOR+ ment, whereas whites have the worse off than are monoracial white biracial group, with that lowest educational attainment of Californians (17% versus 14% multiracial group experiencing the three groups. poor) (Figure 6).5 higher rates of poverty than either

13 Public Policy Institute of California

California Counts California’s Multiracial Population

ulation is slightly more liberal Figure 6. Poverty Rates, Census 2000 and more likely to be registered as Democrats and Independents 30 and less likely than the mono- racial population to be registered 25 Republicans.

20 However, we generally found that multiracial groups have more 15 in common with their monoracial

Percentage peers than with other multiracial 10 groups. The variation between the 5 multiracial groups is almost as large as the variation between the 0 monoracial groups. For example, SOR White White Asian White AIAN White Black most multiracial individuals are

MultiracialSingle race SOR+white Asian+white AIAN+white Black+white slightly less likely than their single- SOR, Hispanic White, Hispanic race counterparts to be registered SOR+white, Hispanic Republican (Figure 7).6 Multi- Source: Authors’ tabulations using Census 2000 5% PUMS. racial blacks, however, are more likely than single-race blacks to be Republican. of its monoracial components. As than 2,000 randomly selected Multiracial Californians appear noted above, this group is more California adults to learn about to be less engaged in the political likely than either monoracial the political views and activities process in general, although these whites or SORs to be foreign-born. of the multiracial population differences are very slight. Adults (Baldassare, 2000). In addition to who have ever identified as multi- identifying themselves as belonging racial are slightly less likely than to one of the standard race cate- those who have not to be regis- Political Attitudes gories, the survey respondents tered to vote. Multiracial adults are and Participation were asked to indicate if they had less likely than adults who have ever identified as “mixed race.” never identified as multiracial to everal groups, such as Arab- Approximately 14 percent indicated say that they vote “always” or SAmericans, advocated for that they had. We also know which “nearly always.”7 recognition in Census 2000, but race they primarily identify with In general, rather than being a without a doubt multiracial groups when asked to choose just one. politically engaged force, multiple- were the most successful. Does Using the survey results to race individuals are not more likely this recognition reflect an engaged compare multiracial adults to to vote, nor do they appear to and highly political population? monoracial adults, we found few share the same political leanings. Since the census does not differences in political attitudes Although they are slightly more include questions about political and participation. Those small likely to be liberal leaning overall, participation or beliefs, we turned differences that do stand out this is not universally true (e.g., to a telephone survey of more suggest that the multiracial pop- multiracial blacks).

14 Public Policy Institute of California

California Counts California’s Multiracial Population

the proportion of multiracial Figure 7. Percentage Registered as Republican Among whites among all whites is small: Single-Race and Multiracial Individuals 4 percent in California; and the 40 characteristics of these multiracial 35 whites are not so different from 30 the characteristics of monoracial 25 * whites. 20 However, for American Indians, + 15 the problem is severe. The most Percentage * 10 important and basic temporal mea- 5 sure of any population—change 0 in its size—is in serious doubt. White White Asian only Asian Black only Black Hispanic Hispanic only multiracial multiracial multiracial only multiracial Census 2000 counted 178,984 White Asian Black Hispanic non-Hispanics who identified only as American Indian, and 383,197 who identified as either American Source: Authors’ calculations from Baldassare (2000). Notes: Simulation controls for age, education, and nativity. Noncitizens are excluded. Indian alone or AIAN multiracial. * Significant difference at the 1 percent level between the group and white. + Significant difference at the 1 percent level within the group between mixed and single race. Census 1990 identified 184,065 non-Hispanic American Indians (Census 1990 Summary Tape File 1). The higher count suggests that Implications Because racial categorizations the American Indian population between Census 1990 and Census more than doubled between 1990 ith the advent of the new 2000 are not consistent (with the and 2000, whereas the lower count Wrace question on the census, addition of the ability to choose suggests a declining population we are finally able to gather data more than one race in 2000), ana- among American Indians. Charac- which allows individuals to state lyzing changes in racial popula- teristics of the American Indian their identity in more detail and tions from one census to the next population in California, including thus helps us understand the com- is difficult. The same problem measures of economic well-being, plexities of racial identification. exists in other datasets, such as the depend on whether we choose Changes in census racial and ethnic Current Population Survey, which to include multiracial American categories have important impli- modified its racial and ethnic cate- Indians, who tend to be better cations for our understanding of gories to allow more than one race educated and have higher wages population change and issues asso- beginning in 2003. The severity than monoracial American Indians. ciated with population counts. of the problem depends on how The ability to choose more than These issues include redistricting, many people choose to identify as one race has made it very difficult civil rights enforcement, and the being of more than one race and to study changes over time for the determination of population-based the extent to which these people American Indian population of rates (including vital statistics, such differ from monoracial respon- California. as mortality and fertility rates). In dents. For most groups, the prob- Demographers have devised this section, we discuss some of lem is neither too extensive nor several alternative methods of these implications. too severe (Table 6). For example, bridging racial categories across

15 Public Policy Institute of California

California Counts California’s Multiracial Population

censuses to address these counting Table 6. Race and Ethnicity Alone and in Multiracial problems. These methods range from counting multiracial people Combinations, Census 2000 as fractions of persons for each Non-Hispanic Hispanic race checked—for example, a White alone 15,816,790 4,353,269 black+white individual would add White alone and in combination 16,538,491 4,952,482 one-half to the black population and one-half to the white popu- Black alone 2,181,926 81,956 lation—to counting multiracial Black alone and in combination 2,370,367 142,674 people as whole individuals for each race checked (essentially Asian alone 3,648,860 48,653 double counting for biracial peo- Asian alone and in combination 4,030,025 125,660 ple). For purposes of civil rights monitoring and enforcement, AIAN alone 178,984 154,362 the OMB (2002) has issued the AIAN alone and in combination 383,197 244,365 following rules: • Responses in the five single-race SOR alone 71,681 5,610,560 categories are not allocated. SOR alone and in combination 368,168 6,207,457 • Responses that combine one minority race and white are allo- Source: Census 2000 SF1, Tables P8 and P10. cated to the minority race. • Responses that include two or more minority races are allocated health researchers use census data rates are derived from a single data as follows: to provide denominators in the source and therefore do not involve – If the enforcement action is calculation of race-specific rates— inconsistent measures of race be- in response to a complaint, including birth rates, death rates, tween numerators and denomina- allocate to the race that the rates of morbidity, crime rates, tors. Estimates of fertility are very complainant alleges the dis- and arrest rates. However, com- different, however, when we com- crimination was based on. bining data from one set of data bine birth records with census data – If the enforcement action with census data can be problem- to estimate total fertility rates in requires assessing disparate atic, especially for multiracial indi- California. These results suggest impact or discriminatory viduals for whom the reporting of that AIAN+white women in Cali- patterns, analyze the pat- racial identity may not be consis- fornia have astonishingly low levels terns based on alternative tent (for a longer discussion, see of fertility, with a total fertility allocations to each of the Johnson and Hayes, 2004). For rate of 1.0 (lower than the total minority groups. example, estimates of fertility for fertility rate for women in any The ability to check more than multiracial women vary depend- country in the world). Of course, one race also raises challenges in ing on the datasets used. Adminis- both results cannot be correct. estimating population-based rates trative data from California birth It is probable that many women because people may be inconsis- records suggest that AIAN+white who identified as AIAN+white tent in their use of biracial identi- women have substantially higher on Census 2000 did not identify ties. Many social scientists and fertility rates than whites. These themselves that way when they

16 Public Policy Institute of California

California Counts California’s Multiracial Population

gave birth in California. Thus, between blacks and whites, while Our reading of Census combining administrative data on narrowing, is still large. With respect 2000 results suggests births with census data on popu- to wages, for example, blacks still that monoracial lations understates fertility rates earn significantly less than whites for this group. Similarly, death even after controlling for age, edu- categories are alive rates based on administrative vital cation, and occupation (Reed and and well, at least for records on deaths and census Cheng, 2003). Nonetheless, in the the time being. counts are unrealistically low for post civil rights era, racial dispari- multiracial populations, with too ties have been reduced. The collec- few death records identifying indi- tion of racial data by government more familiar with their choices viduals as multiracial relative to agencies has allowed us to measure in identifying their races, they may census counts. this improvement. Indeed, the become more likely to use the Supreme Court has affirmed the option to pick more than one. salience of these racial categories, Furthermore, the mere presence Conclusion and census results show that the of these new racial-choice options vast majority of Californians iden- may increase interest in racial her- here are two extreme views of tify with a single race. OMB has itage and ancestry. Certainly, con- Tthe future of racial categoriza- issued guidelines for civil rights tinued increases in intermarriage tion in the United States. In one monitoring that largely rely on and births to parents of different view, racial categories would no monoracial categories. races could greatly expand the longer exist—the Census Bureau Over time, it is hard to project number of multiracial children and other government agencies what might happen to the multi- over time. As their numbers would cease to collect and dissem- racial population—will it lead to increase, a greater proportion of inate data on the basis of race. a new multiracial identity that is young people born to parents of Proponents of this view argue that distinct from underlying mono- different races may choose to retain racial categorization has no scien- racial component groups? One cer- a multiracial identity as they grow tific basis and serves only to pro- tainty is that racial categories will older. It is possible, however, that mote divisive identity politics. In change. Since 1790, no more than the proportion of the population California, this view of the future three consecutive censuses have used identified as being of more than was put to the voters in a statewide the same racial categories. As cen- one race could decrease with time. proposition that the voters rejected sus categories and OMB require- Census questions on race and eth- in 2003. In the other extreme view, ments change, other agencies that nicity could revert back to their there will be an ever-expanding collect data will also change their earlier one-race options. Further- list of groups seeking and gaining racial reporting requirements. more, future state propositions inclusion in the federal statistical Our reading of Census 2000 could be successful in limiting the system (Prewitt, 2002). Census results suggests that monoracial state’s options to collect race data. 2000, the first ever to allow respon- categories are alive and well, at Only the future will tell how dents to identify as being of more least for the time being. Despite rapidly the population who identi- than one race, could be seen as a increasing diversity and intermar- fies as multiracial will grow, and step toward this second future. riage, only 5 percent of Californi- whether it will coalesce into a more Race still matters in California. ans were identified as multiracial. uniform identity shared across In particular, the economic divide However, as individuals become particular racial combinations.◆

17 Public Policy Institute of California

California Counts California’s Multiracial Population

Notes Bibliography

1 This issue of California Counts is part of a Baldassare, Mark, Californians and Their Prewitt, Kenneth, “Race in the 2000 Census: larger research effort examining the multiracial Government, PPIC Statewide Survey, San A Turning Point,” in Joel Perlmann and population of the United States (Tafoya, John- Francisco, California, August 2000. Mary C. Waters (eds.), The New Race Ques- son, and Hill, 2004) and the multiracial iden- tion: How the Census Counts Multiracial Indi- tity of children (Johnson, Hill, and Tafoya, 2004). Bennett, Claudette, Jorge del Pinal, and Art viduals, Russell Sage Foundation, New York, Cresce, “Exploring the Consistency of Race 2002. 2 Collection of data on Americans of Hispanic Reporting in Census 2000 and the Census ethnicity is required to fulfill the obligations Quality Survey,” presentation at the Annual Reed, Deborah, “Recent Trends in Income of Public Law 94-311, which was passed in Meeting of the Population Association of and Poverty,” California Counts: Population 1976 to remedy discrimination against those America, Minneapolis, Minnesota, 2003. Trends and Profiles, Vol. 5, No. 3, February of Hispanic origin. 2004. Bentley, Michael, Tracy Mattingly, Christine 3 The rule employed by the Census Bureau is Hough, and Claudette Bennett, “Census Reed, Deborah, and Jennifer Cheng, Racial that if at least 90 percent of an ethnic group Quality Survey to Evaluate Responses to the and Ethnic Wage Gaps in the California Labor identifies with a single race, any who indicate Census 2000 Question on Race: An Introduc- Market, Public Policy Institute of California, an ethnicity only, without a race, may be tion to the Data,” U.S. Census Bureau, Cen- San Francisco, California, 2003. recoded to the race used by 90 percent of that sus 2000 Evaluation B.3, Washington, D.C., ethnic group. For example, a write-in response April 3, 2003. Sweet, E., Decennial Statistical Studies Divi- of “Jamaican” would be coded as black or sion, U.S. Census Bureau, 1990 REX Memo- African American because 90 percent of California Health Interview Survey, “CHIS randum Series # PP-9, Washington, D.C., Jamaicans identify as black, and a response of Adult Public Use File, Release 2” (computer 1994. “Mexican” would be coded as Some Other file), UCLA Center for Health Policy Tafoya, Sonya M., “Check One or More … Race because fewer than 90 percent of Mexi- Research, Los Angeles, California, 2002. Mixed Race and Ethnicity in California,” cans identify as white. This rule explains the California Counts: Population Trends and Pro- preponderance of Hispanics in the SOR cate- Grieco, Elizabeth M., and Rachel C. Cassidy, files, Vol. 1, No. 2, January 2000. gory, a topic to be addressed below. “Overview of Race and Hispanic Origin: 2000,” Census 2000 Brief, U.S. Census Tafoya, Sonya M., “Latinos and Racial Identi- 4 The relative youth of the multiracial popula- Bureau, Washington D.C., March 2001. fication in California,” California Counts: tion raises an important caveat about census Population Trends and Profiles, Vol. 4, No. 4, data collection. Data are collected by house- Johnson, Hans P., and Joseph M. Hayes, May 2003. hold. Generally one person in each home pro- “The Demographics of Mortality in Califor- vides information on all household members. nia,” California Counts: Population Trends and Profiles, Vol. 5, No. 4, May 2004. Tafoya, Sonya M., Hans P. Johnson, and Usually a parent, and more than likely a Laura E. Hill, “Who Chooses to Choose mother, completes the census form for minors Johnson, Hans P., Laura E. Hill, and Sonya Two?” Russell Sage Foundation, New York, in the household (Sweet, 1994). In this analy- M. Tafoya, “Which Parents Choose to New York, and Population Reference Bureau, sis, it is impossible to know whether the child Choose Two? Multiracial Identity of Chil- Washington, D.C., forthcoming, 2004. perceives himself or herself in the same way dren,” presentation at the Annual Meeting of he or she is perceived by the adult respondent the Population Association of America, Boston, Tucker, Clyde, Ruth McKay, Brian Kojetin, charged with providing the information. Massachusetts, 2004. Roderick Harrison, Manuel de la Puente, Linda Stinson, and Ed Robison, “Testing 5 The census calculates poverty rates for families. Jones, Nicholas A., and Amy Symens Smith, Methods of Collecting Racial and Ethnic The rates reported in the figure reflect family- “The Two or More Races Population: 2000,” Information: Results of the Current Popula- level poverty rates assigned to individuals. Census 2000 Brief, U.S. Census Bureau, tion Survey Supplement on Race and Eth- nicity,” Statistical Notes, No. 40, Bureau of 6 These findings exclude noncitizens and con- Washington, D.C., 2001. Labor Statistics, 1996. trolled for the age, education level, and nativity Office of Management and Budget, “Guid- of respondents. The regression has indicator ance on Aggregation and Allocation of Data variables for each monoracial group and each on Race for Use in Civil Rights Monitoring multiracial combination (mixed-race black, and Enforcement,” OMB Bulletin 00-02, mixed-race Asian, etc.). 2002, available at http://www.whitehouse. 7 These findings exclude those who are not gov/omb/bulletins/b00-02.html. registered, and control for age, education level, and nativity of respondents. The regres- sion has indicator variables for each monora- cial group and each multiracial combination (mixed-race black, mixed-race Asian, etc.).

18 Public Policy Institute of California

California Counts California’s Multiracial Population

Board of Directors Raymond L. Watson, Chair David W. Lyon Constance L. Rice Vice Chairman of the Board Emeritus President and CEO Co-Director The Irvine Company Public Policy Institute of California The Advancement Project Edward K. Hamilton Vilma S. Martinez Thomas C. Sutton Chairman Partner Chairman & CEO Hamilton, Rabinovitz & Alschuler, Inc. Munger, Tolles & Olson LLP Pacific Life Insurance Company Gary K. Hart Cheryl White Mason Cynthia A. Telles Founder Chief, Civil Liability Management Department of Psychiatry Institute for Educational Reform Office of the City Attorney UCLA School of Medicine California State University, Sacramento Los Angeles, California Carol Whiteside Walter B. Hewlett Arjay Miller President Director Dean Emeritus Great Valley Center Center for Computer Assisted Graduate School of Business Research in the Humanities Stanford University Ki Suh Park Design and Managing Partner Gruen Associates

The Public Policy Institute of California is a private, nonprofit research organization established in 1994 with an endowment from William R. Hewlett. The Institute conducts independent, objective, nonpartisan research on the economic, social, and political issues affecting Californians. The Institute’s goal is to raise public awareness of these issues and give elected representatives and other public officials in California a more informed basis for devel- oping policies and programs. PPIC does not take or support positions on any ballot measure or on any local, state, or federal legislation, nor does it endorse, support, or oppose any political parties or candidates for public office.

PUBLIC POLICY INSTITUTE OF CALIFORNIA 500 Washington Street, Suite 800 • San Francisco, California 94111 Telephone: (415) 291-4400 • Fax: (415) 291-4401 • www.ppic.org

19 Public Policy Institute of California

California Counts California’s Multiracial Population

Recent issues of California Counts POPULATION TRENDS AND PROFILES

The Demographics of Mortality in California Recent Trends in Income and Poverty California’s Newest Immigrants California’s Newest Neighborhoods Latinos and Racial Identification in California Maturity Before Maternity: Teen Birth Rates in California California’s Young Children: Demographic, Social, and Economic Conditions

are available free of charge on PPIC’s website www.ppic.org

PUBLIC POLICY INSTITUTE OF CALIFORNIA NON-PROFIT ORG. 500 Washington Street, Suite 800 U.S. POSTAGE San Francisco, California 94111 PAID BRISBANE, CA PERMIT #83

In This Issue California’s Multiracial Population