The socioeconomic status of Asian Brazilians in 1980

MISSING IMAGE

Material Information

Title:
The socioeconomic status of Asian Brazilians in 1980 a comparation of Asians, whites and Afro-Brazilians
Physical Description:
xiv, 277 leaves : ill. ; 29 cm.
Language:
English
Publication Date:

Subjects

Subjects / Keywords:
Brazilians -- Social conditions -- Brazil   ( lcsh )
Brazilians -- Economic conditions -- Brazil   ( lcsh )
Asians -- Social conditions -- Brazil   ( lcsh )
Asians -- Economics conditions -- Brazil   ( lcsh )
Brazilians -- Statistics -- Brazil   ( lcsh )
Whites -- Social conditions -- Brazil   ( lcsh )
Whites -- Economic conditions -- Brazil   ( lcsh )
Blacks -- Social conditions -- Brazil   ( lcsh )
Blacks -- Economic conditions -- Brazil   ( lcsh )
Anthropology thesis Ph.D
Dissertations, Academic -- Anthropology -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1994.
Bibliography:
Includes bibliographical references (leaves 266-276).
Statement of Responsibility:
by Jirimutu.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001981106
oclc - 31913502
notis - AKF7993
System ID:
AA00002065:00001

Full Text









THE SOCIOECONOMIC STATUS OF ASIAN BRAZILIANS IN 1980:
A COMPARATION OF ASIANS, WHITES AND AFRO-BRAZILIANS














By
JIRIMUTU


DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY














ACKNOWLEDGEMENTS


I wish to thank Dr. H. Russell Bernard, my advisor, for his persistent
teachings in the scientific approach to anthropological inquiries and


positivistic attitude toward human knowledge.


I also appreciate his generous


support, academic and otherwise, and his genuine understanding and belief


in me during the course of my graduate studies.
have been entirely impossible if Dr. Charles H.


This dissertation would


Vood had not sparked my


interest in demographic studies when I took a seminar on population with


him.


Throughout my dissertation research, I have benefited tremendously


from his breadth of knowledge, mastery of computer skills and data analysis


techniques, and friendship.


He has my deepest gratitude for inentoring me in


every way possible, even after he moved to another university.


thanks go to Dr. Paul J.


My sincere


Maganrella for his academic support and personal


friendship, both of which are very important for a foreign graduate student


like me.


I thank Dr. Paul L. Doughty for serving on my committee and


offering his wisdom.


I appreciate Dr. Barbara Ann Zsembik for joining my


committee and offering her expertise as a demographer.


I am extremely


grateful to the Wenner-Gren Foundation for Anthropological Research for
providing me with financial support for the first three years of my graduate


school.


Finally, I sincerely thank my wife, Mingxin Zhang, and my son,





















. i


TABLE OF CONTENTS

pAag
ACKNOWLEDGEMENTS .........................................................................................ii


LIST OF TABLES..


LIST OF FIGURES


ABSTRACT


CHAPTERS


INTRODUCTION


Asian Immigrants in the United States...............
Research Design ...... ......................... .....................
Asian Immigrants in Brazil......................................

HISTORICAL OVERVIEW OF THE JAPANESE


EXPERIENCE IN BRAZIL.


Historical Background for the Japanese Migration to Brazil ................14
Japanese Immigration to Brazil ................................................................16
Social Characteristics and Social Mobility of the


Japanese Immigrants
Summary .......................


....... ................ ...................... .... ........4.... 1


FERTILITY DIFFERENTIALS AMONG ASIANS,
AND AFRO-BRAZILIANS ..........................................


WHITES
. .. ....... ...... ...........48


A Brief Review of Literature on Fertility Studies ..................................48
Fertility Differentials among Ethnic/Racial Groups
in M odern States ................................................ .. .... .. .......... ........................ 53


-- Fertility Differentials Among Asians,
Afro-Brazilians ..............................


Whites and


Summary


....... ...................................................................................................xii


0"...'...........0.........................O................... ......00.......0...0.* ..4 ..1


.......... .......o..................................... .B............... 14


__ _II ___~









Child Mortality Differentials and Life Expectancy


by Color Group ........................
Sum m ar y ..........................................


EDUCATIONAL ATTAINMENT OF


ASIANS,


WHITES


AND AFRO-BRAZILIANS .....................................

School Attendance Rate of Children Ages 6-16 ..
Educational Attainment of Men Ages 18-65 ......


Educational Attainment of Women Ages


18-65


Summary ....................................................................................

OCCUPATIONAL PROFILE OF ASIANS, WHITES AND


AFRO-BRAZILIANS ...


Occupational Profile of Men Ages 18-65 .....
Occupational Profile of Women Ages 18-65
Summary ...........................................................


MEAN INCOME OF
AFRO-BRAZILIANS


ASIANS, WHITES AND


Mean Monthly Income of Men Ages 18-65


Mean Monthly Income of Women Ages
Summary .....................................................


18-65


..188
..205
..221


SUMMARY


AND CONCLUSION


APPENDICES

A. BRAZILIAN RACIAL CATEGORIES AND THE CENSUSES.........247

B INDIRECT MEASURES OF CHILD MORTALITY .................................254

C LOGISTIC REGRESSION WITH SCHOOL ATTENDANCE
RATE OF CHILDREN AGES 6-16 AS THE DEPENDENT
VARIABLE, METROPOLITAN SAO PAULO, BRAZIL, 1980..........258

D OCCUPATIONAL CATEGORIES IN THE 1980 CENSUS.................264


...........................................................2


'I).. .... ....tQtt O........O.i... .'.....''''.''.. 8


t......................................89


....................................91


.o......o... ................1 .32


.... ....................... .. .......o..........187













LIST OF TABLES


Table

1.1


1.2


Page
Distribution of Amarelos Ages 15-65 by Place of Birth and
National Origin, Metropolitan Sao Paulo, Brazil (1980)...............13


Racial Composition of Brazil's Population,


Industry Distribution of Brazilian Males Aged 10 and


Over by Color, 1950..


........................ .......23


Employment Status of Brazilian males Aged 10 and Over


for All Industries and Agriculture by Color, 1950.


.................24


Proportion of Farmers Among Japanese Immigrants and
Descendants Aged 10 and Over in Labor Force by Sex,
Brazil, 1958..............................................................................


................25


Occupational Distribution of Japanese Immigrants and
Descendants Aged 10 and Over in Labor Force,


Brazil


1958..................................................................


Japanese Immigrants and Descendants Aged 10 and Over
in Labor Force by Industry, Brazil, 1958................................


.................28


A Comparison of Occupational Status of Prewar and
Postwar Non-Farming Japanese Immigrants....................................30

Agricultural Production of Japanese Brazilians in Sao Paulo


and Brazil by Crop, 1958.................


Japanese Immigrants and Descendants Aged 7 and Over
by Level of Education and Residence, 1958...................


...... ............36


Marital Status of the Japanese Population in Brazil


.............................27


............32


1940-1980 ........................13










Proportion of Traditional Families Among Japanese
Farmers and Non-Farmers in Brazil by Value of


Property Owned,


Mean Children Ever Born to Women of 15-49 Years of Age
By Age Group, Metropolitan Sao Paulo, Brazil (1980)...................58


Mean Children Ever Born to Women of 15-49 Years of Age
By Color Group, Metropolitan Sao Paulo, Brazil (1980)................59

Mean Children Ever Born to Women of 15-49 Years of Age


By Age and Color Groups, Metropolitan Sao Paulo,
Brazil (1980).......................................................................


Mean Children Ever Born to Women of 15-49 Years of Age
By Education, Age and Color Groups, Metropolitan


Sao Paulo, Brazil (1980)


Mean Children Ever Born to Women of 15-49 Years of Age
By Income, Age and Color Groups, Metropolitan


Sao Paulo, Brazil (1980)


Mean Children Ever Born to Women of 15-49 Years of Age
By Residence, Age and Color Groups, Metropolitan


Sao Paulo, Brazil (1980)


Mean Children Ever Born to Women of 15-49 Years of Age
By Residence, Education and Color Groups,


Metropolitan Sio Paulo,


Brazil (1980).........


Mean Children Ever Born to Women of 15-49 Years of Age
By Residence, Income and Color Groups,


Metropolitan S&o Paulo, Brazil (1980)............


Children Ever Born to Women Aged


Age,


20-49 Regressed on


Residence, Education, Income and Color...................................71


Social Indicators by Color Group, Metropolitan Sio Paulo,
Brazil (1980) ............................................................................


1958..... ........... ...............................


.....................62


S.......................................................... ...............63


............................................... ........ .............65


...... .... ................'. ...........66


......................................68


............82


..................... 40


.....................60









Metropolitan Sao Paulo, Brazil (1980) ...................................................87

Number of Children Ages 6-16 and the Percent in School


by Age, Metropolitan Sao Paulo, Brazil (1980)....


................... .........93


Number of Children Ages 6-16 and the Percent in School
by Color Group, Metropolitan Sho Paulo, Brazil (1980).................93

Distribution of Children Ages 6-16 and the Percent in School
by Income Level, Metropolitan Sao Paulo, Brazil (1980)...............94

Children Ages 6-16 and the Percent in School by Residence,


Metropolitan Sao Paulo,


Brazil (1980)............


Number of Children Ages 6-16 and the Percent in School


by Parents' Education,


Metropolitan Sao Paulo,


.............................96


Number of Children Ages 6-16 and the Percent in School


Sex,


Metropolitan S&o Paulo, Brazil (1980)


..................97


In-School Rate of Children Ages 6-16 and the Percent
in School by Color Groups, Metropolitan Sao Paulo,


Brazil (1980)............................ .......................


.... .... ................. .........98


In-School Rate of Children Ages 6-16 by Income and Color,


Metropolitan Sao Paulo,


In-School Rate of Children Ages 6-16 by Region,


Income


and Color, Metropolitan Sao Paulo, Brazil (1980)............................102


5.10


Logistic Regression of In-School Rate of Children Ages


6-16 on Mother's and Father's
Income, Residence and Color b


Education, Household
'y Age, Metropolitan


Sao Paulo, Brazil (1980)...................................... .... ............... ...........1

Mean Years of Schooling for Men Ages 18-65 by Color Group,


Metropolitan S8o Paulo,


Brazil (1980).......


........114


Mean Years of Schooling for Men Ages 18-65 by Age Group,


Ilo.rn-nn- a C,- PT.iii1,n Ir-,-1I (1Qfn\


................95


Brazil (1980).............................................................


Brazil (1980)........ ..................... .....................100


11A









Metropolitan S o Paulo,


Brazil (1980) ... .... .... .. ............ ...... .............116


Mean Years of Schooling for Men Ages 18-65 by Residence


and Color, Metropolitan Sao Paulo, Brazil (1980)


Mean


Years of Schooling for Men Ages 18-65 by Income and


Color, Metropolitan Sho Paulo,


Mean


Me


Brazil (1980)..


.........118


Years of Schooling for Women Ages 18-65 by Color Group,
tropolitan Sio Paulo, Brazil (1980)................................. ...............120


A


Mean


Years of Schooling for Women Ages 18-65 by


Metropolitan Sao Paulo, Brazil (1980)...


Age Group,
.........................121


5.19


Mean


Years of Schooling for Women Ages 18-65 by Residence,


Metropolitan Sio Paulo,


5.20


Brazil (1980)...


.....121


Mean Years of Schooling for Women Ages 18-65 by Age
and Color, Metropolitan Sao Paulo, Brazil (1980)........


Mean


Years of Schooling for Women Ages 18-65 by Residence


and Color, Metropolitan Sao Paulo,


Mean


Brazil (1980).


.......126


Years of Schooling for Women Ages 18-65 by Income


and Color, Metropolitan Sao Paulo, Brazil (1980) ........................126


Occupational Distribution of Men Ages 18-65 by Color Group,


Metropolitan Sao Paulo, Brazil (1980)........


........134


Top Five Occupations in the Category of Unskilled/Personal
Service by Color, Metropolitan Sho Paulo, Brazil (1980).....


Occupational Distribution of Men Ages 18-65 by Residence,


Metropolitan Sao Paulo, Brazil (1980)


......0.............. ..137


Occupational Distribution of Men Ages 18-65 by Age Group,
Metropolitan Sio Paulo, Brazil (1980)...................................


..........138


Occupational Distribution of Men Ages 18-65 by Income,


Metropolitan Sao Paulo, Brazil (1980)


................. ............ ...... 140


.................117









Paulo, Brazil (1980)............................................................. ....... ..... ...144

Occupational Distribution of Men Ages 18-65 by Residence
and Color, Metropolitan Sao Paulo, Brazil (1980).........................145

Proportion of White vs. Blue Collar Occupations of Men
Ages 18-65 by Age and Color, Metropolitan Sfo Paulo,


Brazil (1980)...................


.............147


6.10


Occupational Distribution of Men Ages 18-65 by Age
and Color, Metropolitan Sio Paulo, Brazil (1980)........................... 149


Occupational Distribution of Men Ages 18-65 by Income
and Color, Metropolitan Sio Paulo, Brazil (1980)............................ 151

Occupational Distribution of Men Ages 18-65 by Education
and Color, Metropolitan Sao Paulo, Brazil (1980).........................157


Occupational Distribution of Women Ages 18-65 by Color


Group, Metropolitan Sao Paulo, Brazil (1980)


..... ........ ...... .... .. .1


Occupational Distribution of Women Ages 18-65 by Residence,


Metropolitan S&o Paulo, Brazil (1980)....


Occupational Distribution of Women Ages 18-65 by
Metropolitan S&o Paulo, Brazil (1980).....................


Occupational Distribution of Women Ages


.........................160

Age Group,
.........................162


18-65 by Income


Level, Metropolitan S&o Paulo, Brazil (1980)...


............ ... .............164


Occupational Distribution of Women Ages 18-65 by Education,


Metropolitan S&o Paulo, Brazil (1980).............


.....166


6.18


Occupational Distribution of Women Ages 18-65 by Residence


and Color, Metropolitan Sio Paulo, Brazil (1980).


.......168


Occupational Distribution of Women Ages 18-65 by


Age


and Color, Metropolitan Sao Paulo, Brazil (1980).........................171


Occupational Distribution of Women Ages 18-65 by Income
and 'nolnr Mfatrnrnlifan <3 Pmin Rr2711 (1QRNf 171


6.14









Metropolitan So Paulo, Brazil (1980) ................................ .......189

Mean Monthly Income of Men Ages 18-65 by Age Group,


Metropolitan Sao Paulo, Brazil (1980)........


...............190


Mean Monthly Income of Men Ages 18-65 by Residence,


...........................................190


Metropolitan Sao Paulo, Brazil (1980)..


Mean Monthly Income of Men Ages 18-65 by Education,


Metropolitan Sao Paulo, Brazil (1980)...


............. ............................. 191


Mean Monthly Income of Men Ages 18-65 by Occupation,
Metropolitan Sio Paulo, Brazil (1980)......................................... .......193

Mean Monthly Income of Men Ages 18-65 by Age
and Color, Metropolitan Sio Paulo, Brazil (1980)........................194

Mean Monthly Income of Men Ages 18-65 by Residence
and Color, Metropolitan S&o Paulo, Brazil (1980)........................ 196

Mean Monthly Income of Men Ages 18-65 by Education
and Color, Metropolitan Sao Paulo, Brazil (1980)........................197

Mean Monthly Income of Men Ages 18-65 by Occupation
and Color, Metropolitan Sao Paulo, Brazil (1980)......................... 199


Monthly Income of Men Ages 18-65 Regressed on Age,


Education


, Residence and Color, Metropolitan


SAo Paulo, Brazil (1980)...............


Mean Monthly Income of Women Ages 18-65 by Color Group,
Metropolitan Sao Paulo, Brazil (1980). .......................................


Mean Monthly Income of Women Ages 18-65 by
Metropolitan S&o Paulo, Brazil (1980)..................


Age Group,
..............................206


Mean Monthly Income of Women Ages 18-65 by Residence,


Metropolitan Sao Paulo, Brazil (1980).


Mean Monthly Income of Women Ages 18-65 by Education,


NI a%,-dHar^ cn P.guln fl,.^rwfl 0 0fo\


.....205


..0.0.........................207


*I ItJ


/ r. fn


.............................204









and Color, Metropolitan Sao Paulo, Brazil (1980)........................211


7.17


Mean Monthly Income of Women Ages


18-65 by Residence


and Color, Metropolitan Sao Paulo, Brazil (1980).... ......................212


Mean Monthly Income of Women Ages 18-65 by Education
and Color, Metropolitan S&o Paulo, Brazil (1980).........................213


Mean Monthly Income of Women Ages 18-65 by Occupation


and Color, Metropolitan Sao Paulo, Brazil (1980).......


Monthly Income of Women Ages


18-65 Regressed on


Age, Education, Residence and Color, Metropolitan
Slo Paulo, Brazil (1980).......................................................................220


.. ...............2












LIST OF FIGURES


Figure


Page


In-School Rate of Children Ages 6-16 by Age and Color Groups,


Metropolitan SAo Paulo, Brazil (1980)..............................


.......99


In-School Rate of Urban Children Ages 6-16 by Income and
Color Group, Metropolitan Sao Paulo, Brazil (1980).........


In-School Rate of Rural Children Ages 6-16 by Income and
Color Group, Metropolitan Slo Paulo, Brazil (1980)........


Effects of Father'


and Mother's Education on Children'


In-School Rate, Metropolitan Sao Paulo, Brazil (1980).


Effects of Household Income and Urban Residency
on Children's In-School Rate, Metropolitan
Sio Paulo, Brazil (1980).......................................................................111

The Odds of Afro-Brazilian and Asian Children Being in
School Against Those of White Children,
Metropolitan Sao Paulo, Brazil (1980)................................................112


Mean


Years of Schooling by Sex and Color Group,


Metropolitan Sio Paulo, Brazil (1980)


Mean


Years of Schooling by Sex and Age Group,


Metropolitan Sao Paulo, Brazil (1980)


Mean


Years of Schooling by Sex and Residence,


Metropolitan Sao Paulo, Brazil (1980) ............... ............... .....................123


.............................122


.............................................122













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


THE SOCIOECONOMIC STATUS OF ASIAN BRAZILIANS IN 1980:


A COMPARISON OF


ASIANS,


WHITES AND


AFRO-BRAZILIANS


Jirimutu
April 1994


Chairman:


Major Department:


I use the 3%


H. Russell Bernard


Anthropology


sample data of Metropolitan Sao Paulo from the 1980


Brazilian Census to examine the socioeconomic standing of Asian Brazilians,


relative to whites and Afro-Brazilians in Brazil.


Operationally


socioeconomic standings of the three color groups are measured in terms of


fertility level,


child mortality level and the life expectancy rate associated with


it, educational attainment (school attendance rate of children ages 6-16 and
years of school completed by men and women age 18-65), occupational profile


and mean monthly income of men and women ages 18-65.


The results of


these measurements indicate that in 1980 Asian Brazilians lead both whites









in Brazil


, the presence of ethnic enclaves and economies, ownership of small


business, continued heavy investment in education through several
generations, the family characteristics that facilitate stability and capital

accumulation, and the cultural values, such as hard work, industriousness,


emphasis on education,


obligation and loyalty to family and kin group.


This


shows that Asian immigrants in Brazil have experienced similar, if not more,
success in upward social mobility as have Asian immigrants in the United

States.













CHAPTER 1
INTRODUCTION


Asian Immigrants in the United States


Asian immigrants in the United States have long attracted the

attention of social scientists because it is widely recognized that they have


done very well,


over time


, compared to many other immigrant groups.


They


are said to have achieved parity with or even surpassed the majority whites


in socioeconomic standing (Bell,


1985; Bonacich & Modell,


1980


Chiswick,


1980; Hirschman, 1983; Hirschman & Wong, 1981


Hirschman & Wong, 1986; Jiobu,


Hirschman & Wong, 1984;


1976; Jiobu, 1990; Kitano, 1974; Kitano &


Daniels, 1988; Montero,


1981


Montero &


Tsukashima


1977


Nee & Sanders,


1985; Nee & Wong, 1985; Petersen, 1971; Ro,

Asian Americans have been called a


se, 1985; Wong, 1980; Wong, 1982).

"model minority" (Newsweek,


1982) and "America's


super minority" (Ramirez, 1986),


and hailed as


"America's


greatest success story" (Bell,


1985).


Indeed


, Asian Americans,


especially Japanese Americans and Chinese Americans,


have achieved great


success in terms of labor force participation, income and education.


In 1979,


95% of Asian Americans (the six largest groups within Asian Americans
which include Chinese, Filipino, Japanese, Asian Indians, Korean and

Vietnamese) had a median family income of $23.600. compared to the average









advances in income and education for Asian Americans.


Their median


family income in 1989 was $41,583 compared with the national average of


$35,225, and 38% of Asian Americans had graduated with a bachelor's


degree


or higher, compared with 20% of the total population (Bureau of the Census,
1993).


However


, the Asian success story has been exaggerated to some extent


because the statistics on median family income does not reflect the entire


picture of the socioeconomic status of Asian Americans.


have noted


As many researchers


, the higher median family income of Asian Americans is mainly


due to their larger average family size (3.8 persons for Asian American


families vs. 3.2 persons for all U


families in 1989),


higher proportion of


families with three or more workers (19.8% for Asian American vs.


the total population),


13% for


geographical locations (Asian Americans are highly


concentrated in California and New York, where the average income is
higher, relative to the rest of the country), and higher educational attainment.
In fact, the mean personal income of Asian Americans in 1989 was slightly


lower than the national average: Per capital income for


Asian Americans in


1989 was $13,806, compared with the national per capital income of $14,143


(Bureau of the Census, 1993).


Nonetheless, there is no doubt that compared


to many other immigrant groups, most Asian Americans have overcome the
disadvantages that immigrant groups typically confront in the United States.

Social scientists have devoted considerable effort to understanding the


factors that explain the relative success of Asian immigrants.


Some have


stressed the role of


"middleman minorities"


for various Asian groups in the









social mobility (Li,


1977


Lyman, 1977; Nee & Sanders, 1985;


Takaki 1989).


Others have argued that the strength of kinship and family ties, and the


emphasis on education,


hard work and sacrifice for children are mostly


responsible for their success (Kitano, 1969; Newsweek, 1982; Petersen, 1971;


Schwartz


, 1971)


The first two arguments are structural explanations while the third


type is cultural.


The structural arguments mainly examine the relationship


between the minority in question and the society at large in terms of
occupational structure, economic status and the role of ethnic organizations


in the economic, social and political arena.


Cultural arguments either focus


on the cultural characteristics of the minorities themselves or seek
similarities between the cultural values of the dominant society and the

minorities and to attribute the success of minorities to these cultural traits.

Nee and Wong (1985:282) argued that both the cultural and structural

explanations were ahistorical because they "fail to capture the dynamic nature

of immigrant groups as they respond to historical situations and changing


economic structures."


For them, the cultural argument was a form of circular


reasoning and failed to include two important variables that were essential
for the upward mobility of immigrants and their descendants; 1)


"immigrants'


willingness to endure hardship for economic gains" and 2) "the


socioeconomic background at the time of immigration" (1985:283).
They maintained that the cultural characteristics of Asian Americans


reflected the influence of neo-Confucianism,


which emphasized


legitimacy of status attainment through education and membership and


"the









the time of immigration as crucial for "the creation of opportunities for


upward mobility."


Without the necessary human capital to generate


resources, the cultural characteristics of immigrants would have much


smaller impact on their socioeconomic


standing.


Therefore


both the cultural


characteristics and the socioeconomic background of immigrants were
essential in understanding their success in the new country.

Nee and Wong (1985:286) also criticized the structural argument for
"failing to deal with the changing economic condition of the expanding


market economy in North America."


They maintained that the


socioeconomic attainments of immigrant and ethnic groups are the result of
"continuous change and transformations of both cultural attributes and labor


market conditions" (1985:287).


The formation of household production units,


they argued, facilitated the social and economic mobility of Japanese


Americans.


The profit from the household production units in turn served


as the capital for further development of small businesses.


Nee and Wong


particularly stressed the importance of the family bond in the socioeconomic


attainment of


Asian Americans:


Cheap labor generated by household units allowed these ethnic
businesses to be competitive in the dominant society; formation
of family businesses coincided with the development of an


enclave economy,


which opened ethnically controlled avenues


for socioeconomic mobility, and provided a stable environment
for family life and the socialization and education of an
upwardly mobile second generation. (1985:287-288)


Nee and Wong (1985) used a "supply-demand" versvective, which









and the socioeconomic background prior to and after immigration on the
supply side, and put the structural constraints and opportunity structures
created by the development of the capitalist economy on the demand side.
Theories of middleman minorities and of ethnic enclaves are often


applied in the literature of Asian Americans.
ethnic group relations, such as Blalock (1967),


Drawing on earlier works on
Bonacich (1973) and Bonacich


and Modell (1980) argued that certain minorities in multiethnic societies
occupy a middle status between the dominant group and the subordinate


groups, acting as buffers between elites and masses.


These middleman


minorities usually occupy an intermediate niche in the economic system and
tend to concentrate in certain occupations, such as traders, moneylenders and


shopkeepers.


Middleman minorities therefore provide goods and services to


both the elites and the masses.


belonging nowhere,


Because of their unique social position of


they tend to develop strong in-group solidarity and form


their own separate and distinct community.


Two often-cited examples of


middleman minorities are Jews in feudal and early modern Europe and
Chinese in Southeast Asia (also called overseas Chinese) (Bonacich and


Modell


, 1980).


Bonacich and Modell (1980) applied the theory of middleman


minorities to the experience of Japanese Americans.


They argued that


Japanese Americans, particularly the issei, or the first generation, exhibited
many of the traits of a middleman minority; they "formed a highly organized,


internally solidary community,"
nonindustrial family businesses"


"concentrated in self-employment and


and "faced severe hostility from the









"Ethnic enclaves are a distinctive economic formation, characterized by the
spatial concentration of immigrants who organize a variety of enterprises to

serve their own ethnic market and the general population" (Portes and Bach,


1985:203).


The presence of immigrants with sufficient capital to create new


opportunities for economic growth and an extensive division of labor are two


fundamental traits of economic enclaves.


Ethnic enclaves also require a large


number of immigrants with business skills and a large pool of low-wage


labor.


The Cubans in Miami and Koreans in Los Angeles are contemporary


examples of ethnic enclaves (Portes and Manning, 1986).


Some ethnic groups


are highly entrepreneurial, possess capital, and therefore develop ethnic
economies that consist of many small businesses, some of which interface


with the majority economy (Portes and Jensen, 1987).


Within this enclave,


ethnic workers do not have to compete with the majority workers and are
usually not directly subject to discrimination by the dominant group.
Therefore, they can climb the socioeconomic ladder relatively free of racial

and ethnic discrimination.
This does not mean that everyone is equal in an ethnic enclave. On the
contrary, ethnic employers exploit ethnic workers, especially recent arrivals,


and make huge profit from cheap labor.


On the relationship between the


employers and workers in ethnic enclaves, Jiobu (1990:171) stated that "to the


extent that workers rely on enclave employment,


their income


, and by


implication their socioeconomic standing, will be suppressed.


other hand


But on the


, suppressing the income of workers raises the income (and


socioeconomic standing) of ethnic employers."









more comparative research on the fate of Asian immigrants in other


countries, such as Brazil,
Asians (mostly Japanese)


a country that has received a large contingent of

. In contrast to the vast literature on Afro-Brazilians,


the literature, especially recent studies,


on the Asian population in Brazil is


remarkably small


In this dissertation


, I examine whether Asian immigrants


have experienced the same socioeconomic success in Brazil as they have in


the United States.


Specifically,


I compare Asian Brazilians to whites and


Afro-Brazilians in Brazil in terms of quality of life.




Research Design


Dependent Variables


The data for this study are the 3


sample of Metropolitan Sao Paulo


from the 1980 Brazilian Census.


Conceptually


quality of life can be measured


by success or failure at various crucial periods of the life course:


surviving childhood,


Operationally


giving birth,


acquiring an education, finding a job and getting paid.


the dependent variables in this study are fertility level (mean


children ever born to women ages 15-49),


child mortality rate and the life


expectancy rate associated with it, school attendance rate for children ages 6-16


and educational attainment of men and women ages


18-65


, occupational


profile of men and women ages
and women ages 18-65. Taken t


18-65, and mean monthly income of men


together, these variables provide us with a









Color Groups


The 1980 Brazilian census used four categories for racial classification;


branco, pardo, preto and amarelo, or white,


brown


, black and yellow.


In this


study, I use three color groups (whites,
instead of the four in the 1980 census.


Afro-Brazilians and Asian Brazilians),

In other words, I have combined the


census categories of black (preto) and brown (pardo) into a single category


called

"Asian


"Afro-Brazilians,"


Brazilians"


and have replaced "yellow" (amarelo) with the term


or simply "Asians."


My decision of combining the categories of brown and black into a


single category is based on two things; the focus of th


study and the findings


of a number of studies on racial inequalities in Brazil (Hasenbalg, 1985;


Hasenbalg and Huntington,


1982; Lovell,


1989; Silva,


1978; Silva,


1985; Wood,


1990; Wood and Carvalho, 1988; Wood and Lovell


1989


Wood and Lovell,


1992).


First, since the focus of this study is Asian Brazilians, I could have


compared them to the rest of the population as a whole or to all of the racial


categories used in the 1980 census.


In my view, though, the position of Asian


Brazilians in Brazilian society is most clearly shown by comparing them to
whites and Afro-Brazilians since we know from the literature that there are


significant differences among these groups.


Second


, the above studies found


that although there are differences between blacks and mulattos in
socioeconomic standing, they are much closer to one another than to whites


and there are substantial differences between whites and nonwhites.


In other


words


, there is a major dividing line between whites and non-whites.


Ths


Thus,









Independent


Variables


In addition to color group, the most important independent variables


in this study are age,


most chapters,


residence, educational level, and income level.


I treat age as an ordinal variable consisting of three categories


(18-25


, 26-39 and 40-65 years old).


the dependent variables.


This eliminates the general effect of age on


In regression analyses,


age is treated as an interval


variable.


Residence is a dichotomous variable: urban or rural.


Educational attainment is measured by years of schooling completed.


In descriptive analyses, I generally treat this a
of five levels: 1) no schooling at all, 2) one to I


s an ordinal variables consisting

four years of schooling, 3) five to


eight years of schooling, 4) nine to eleven years of schooling and 5) twelve or


more years of schooling.


However, years of schooling is treated as an interval


variable in regression analyses.
Mean monthly income refers to the sum of either household or
individual income from different sources, such as occupation, income in


kind, retirement (social security), rent, gifts,


capital and others,


during the


period of twelve months preceding the census.


1980 (one minimum wage


= 4,150 cruzeiros),


Based on the minimum wage

mean monthly income is


classified into four levels in descriptive analyses:


wage (zero to 4,150 cruzeiros),


51-8,300 cruzeiros),


1) up to one minimum


2) between one and two minimum wages


3) between two and three minimum wages (8,301-


12,450 cruzeiros),


4) above three minimum wages (above


2,450 cruzeiros).


However, in regression analyses mean monthly income is treated as an









Organizations of the Chapters


Chapter


2 provides an historical overview of Japanese migrations to


Brazil and of the Japanese experience in Brazil from their arrival at the turn


of the century to the late 1950s.


Chapter 3 starts with a review of fertility


theories and of racial/ethnic differentials in fertility.


fertility differences by color, age,


I then examine the


educational level, income level and


residence before comparing the fertility differentials among Asians, whites


and Afro-Brazilians, controlling for the other variables.


Finally


I conduct a


multivariate regression analysis to examine the association of fertility and the
other variables.
In Chapter 4, I discuss major determinants of mortality and


racial/ethnic differences in mortality in multiethnic societies.


Then


describe some key socioeconomic indicators of Asian, white and Afro-
Brazilian women and use indirect measures to calculate child mortality rate


for the three color groups.


On the basis of the mortality level for each group, I


calculate the life expectancy rate for each of the three groups and discuss the


implications of these rates.


Finally, I examine the association between the


major socioeconomic indicators and child mortality, using the


Tobit


regression procedure.


Chapter


5 has three sections.


The first section compares Asian,


white


and Afro-Brazilian children ages 6-16 in terms of in-school rate by age,


residence, parents'


educational level and income level.


I then use logistic


regression to measure the effects of these variables on racial differences in









differences in educational attainment.


In the third section, I repeat the same


analysis for the measurement of educational attainment of women ages


18-65.


In Chapter 6, I describe occupational profiles of men and women ages


18-65 separately by color, age, residence,


educational level and income level


and the effects of these variables on racial differences in occupational

distribution.


Chapter


7 describes mean monthly income of men and women


separately by age, residence, occupation and educational level.


I examine the


racial differences in mean income, controlling for the other variables.

In Chapter 8, the concluding chapter, I summarize the main findings of

the study and discuss the implications of my findings in the light of relevant

literature on the experience of Asian immigrants in the United States.




Asian Immigrants in Brazil


Amarelo has been used as one of the four racial categories in the


Brazilian Censuses since 1940


refers.


, and there is little ambiguity as to whom it


Amarelo is designated for people with yellow skin color, who are


either immigrants from Asia or their descendants.


Unlike other racial


categories, there has been very little movement in and out of amarelo.


is probably because of Asians'

intermarriage with other racia

percent of the total population


This


distinct physical features and their lack of

1 groups. Though they comprise less than one

in Brazil, the amarelos are a very stable group


I









records and recent estimates indicate that most of them are of Japanese


descent (Dwyer and Lovell,


1990; Suzuki,


1981


Tsuchida, 1978).


For instance,


of 242,320 amarelos censused in 1940, 99% were Japanese and only


were


Chinese (Tsuchida


,1978).


At the time, Japanese and Chinese were the only


two groups to which the category of amarelo was applied.


By 1980,


not much had changed.


I examined the data on the place of


birth and national origin of Asians (amarelo) aged 15-65, using the 3


sample


data of Sao Paulo from the


1980 Brazilian Census.


The overwhelming


majority of Asians in Brazil are still either Japanese immigrants or their


descendants.


by birth,


The data show that 67


6.8% are naturalized Brazilians and


those who are Brazilian citizens by birth, 92.1


of Asians in the sample are Brazilians


6% are foreign nationals.

were born in Sao Paulo,


followed by


6% from Parana and


from other places.


Meanwhile, of


Asians who were born in foreign countries, 88.1


are from Japan, followed by


from Korea,


countries (see


from China and the remaining


Table 1.1).


2.6% from other


Therefore, we can say with certainty that amarelos,


or "yellow people," are predominantly of Japanese descent, and the Japanese
experience in Sao Paulo constitutes the major part of Asian experience in

Brazil.


The percent of amarel
during the 1940s and 1950s.


in the Brazilian population was


It increased slightly to 0.


steady at 0.6


from the 1960s to the


1980s (see


Table 1.2).


According to the 1980 Brazilian census,


the total


population of amarelos is 673,000. Thri
amarelos as Asian-Brazilians or simply


oughout this stu

Asians, which,


I will refer to


I think, is a more









Paulo and comprised


of the state'


total population (FIBGE 1981).


That is


why I chose the sample data of So Paulo to study

relationships to whites and Afro-Brazilians.


Table 1.1


Asian Brazilians


and their


Distribution of Amarelos Ages


15-65 by Place of Birth and National Origin,


Metropolitan Sio Paulo,


Brazil (1980)


National Origin (


Place of Birth


Brazilian


Naturalized


Foreign


Brazil


Sao Paulo


3,492


Parana
Other


Total


100.0


Foreign
Japan
Korea


3,791


100.0


1,603


China
Other


Total

Total


100.0

100.0


80.7
86.0
45.1
70.8


54.9
29.2
20.9

6.8


----
---


1,819


5,610


67.6


Source: Weighted 3
Census.


sample data of Metropolitan Sao Paulo,


1980 Brazilian


Table 1.2


Racial Composition of Brazil'


Population,


1940-1980


Race 1940 1950 1960 1980
N % N % N% N %

White 26,172 63.5 32,028 61.7 42,838 61.0 64,540 54.2
Brown 8,744 21.2 13,786 26.5 20.706 29.5 46,233 38.8
Black 6,036 14.6 5,692 11.0 6,117 8.7 7,047 5.9
Yellow 242 0.6 329 0.6 483 0.7 673 0.7
N]iccina- Al) N1 1AQ 0 A'7 nA 1 1'7 n A













CHAPTER


HISTORICAL OVERVIEW
OF THE JAPANESE EXPERIENCE IN BRAZIL




Historical Background for the Japanese Migration to Brazil


The overseas migration of Japanese did not start until the Meiji


Restoration of 1868
Meiji Era (1868-191


After that, industrialization and urbanization during the


2) led to massive overseas migration in the late nineteenth


and early twentieth centuries.


Urbanization encroached on agricultural


families and wound up depriving them of access to their land (Ito-Adler,


1987).


Rapid population growth in the rural areas, which exceeded the


industrial growth, also contributed to the massive migration of farmers both


to urban areas in Japan and overseas.


Some analysts (Tsuchida,


1978; Reichl,


1988) argued that the Japanese government considered overseas migration as
a viable option for the increasing problem of surplus rural population.

The first important destinations outside Asia of Japanese emigrants


were Australia (1883),


Hawaii (1885) and Canada (1891).


Reichl (1988:22)


wrote


, "only those Anglo-Saxon countries were sanctioned for emigration


prior to the Russo-Japanese War in 1905 because they


'offered better economic


opportunities than other countries of immigration'


(Tsuchida


, 1978:27).


.I









Sao Paulo and a number of private Japanese emigration companies.
However, Brazil soon became the most important destination for Japanese


immigrants: they became the second largest group (16.8
groups to Brazil during the period from 1924-1941, only


immigrants (33.1%).


%) of all immigrant
after the Portuguese


In fact, by 1938 the Japanese population in Brazil grew to


95,116


, which was the second largest overseas Japanese population,


after that


in Manchuria (233,842), then a colony of Japan (Normano and Gerbi, 1943).
For the period 1950-1955, the Japanese population in Brazil was estimated at


373,000,


making Brazil the country that had the largest Japanese population


outside of Japan,


1959).


followed by the United States (326,376) (Fujii and Smith,


1968, the total number of the Japanese and their descendants in


Brazil was estimated at more than 615,000,


which was 50


of all Japanese


immigrants and their descendants residing in foreign countries.


By then,


United States was a distant second (Sims, 1972).
The serious labor shortage and underpopulation in Brazil in the late
nineteenth and early twentieth centuries were other major factors in the


large-scale emigration of Japanese to Brazil.


Smith (1972:118) cited two major


motivating forces of the Brazilian government for seeking immigrants.
first was "the creation of a small-farming class in the population." The


second was "the ensuring of what Brazil'


upper classes considered an


adequate and cheap labor supply to perform the manual work on the coffee,


cotton


, and sugar plantations of the nation," after the abolition of slavery in


1888.


The Brazilian government preferred Europeans to Asiatic people,









century.


However, several events in Europe and Japan at the turn of the


century had major impacts on the immigration wave to Brazil.
In 1902 the Italian government, in response to reports of mistreatment


of Italian colonos


on plantations in Sao Paulo,


temporarily banned the


subsidized migration of Italian laborers to Brazil. Although Italian laborers
continued to come in small numbers following the ban, they were far too few

to satisfy the growing demand for rural labor on the plantations of SAo Paulo


(Holloway,


1980).


In 1888, Australia prohibited Japanese immigration.
growing anti-Japanese sentiment in North America and t


Agreement"


There was also


between the United States and Japan in 1907 limited


immigration from Japan severely (Reichl,


shortage in Brazil,


1988).


lack of access to Australia and the U


Thus, a severe labor


and Japan'


increasingly overcrowded rural areas created a perfect climate for Japanese


migration to Brazil.


As Normano and Gerbi (1943:45) described it


, "Japan's


land hunger coincided with Brazil'


population hunger."


Japanese Immigration to Brazil


Japanese migration to Brazil can be separated into four time periods,


according to the volume and nature of migration,


and characteristics of


immigrants: 1) 1908-1923,


2) 1924-1941,


3) 1952-1958, 4) 1959-late 1960s.


The Period 1908-1923


"Gentlemen's









their maritime passage.


On the other hand, private Japanese companies were


mostly responsible for the emigration business and the Japanese government


primarily played a coordinating role for most of the time.


The volume of


immigrants during this period was relatively small except the years 1917-1919.


The majority of immigrants were farmers in family units,


Brazilian


as was required by


government.


The first group of Japanese immigrants, consisting of 781 individuals


(158 families),


arrived by ship in the port of Santos,


Sao Paulo, in 1908.


They


came as colonos (contract laborers) under a contract between Japan and the
state of Slo Paulo. During the next fifteen years, Japanese immigrants


continued to come


, though in small numbers.


The total number of Japanese


immigrants from 1908 to 1923 was 32,266, constituting only


of all the


immigrants to Brazil for the time period (Fujii and Smith,


1959).


However,


the period 191


1919 was the peak for the influx of Japanese immigrants,


representing 12.9%,


28.3


Brazil for the three years.


immigrants to Brazi


and 8.4%, respectively, of all the immigrants to

This dramatic increase in the number of


was mostly due to the establishment of the Kaigai Kogyo


-Kabushik (Overseas Development Company), or K.K.K.


Compared to other


Ky> ^ I>


groups, the number of Japanese immigrants was relatively small,


but their


successful beginning was very important to the future of Japanese emigration
to Brazil.


The Period 1924-1941


I *A .









Japanese emigration to Portuguese-speaking America, mainly Brazil, and


away from the earlier destinations in North America.


Both Normano and


Gerbi (1943) and Fujii and Smith (1959) noted that in 1924, the Emigration


Council


, headed by Minister of Foreign Affairs Shidehara, sent a new mission


to South America to explore possible destinations for large-scale emigration.
As a result, the Japanese government decided to concentrate her emigration

effort on Brazil and soon established the Overseas Development Company, a
centralized and highly rationalized management of emigration to Brazil. TJ1

Japanese government also provided subsidies to the company for its

emigration efforts.


Meanwhile


, in 1923 the state of Sao Paulo stopped the policy of giving


subsidies to immigrants from Japan.


The proportion of Japanese immigrants


(of all immigrants to Brazil) increased dramatically from 2.8% in 1924 to


53.2% in 1933, and then steadily decreased to 5.6


II broke out.


in 1941, when World War


This slowdown in the pace of Japanese immigration was also


caused by the Immigration Legislation of 1934 in Brazil, which aimed to
restrict the entry of immigrants annually to two percent of the total entries of


the previous fifty years.


However, the percentage of Japanese among all


immigrants during this period was


previous period,


16.8%, much higher than the


due to the decline of European immigrants.


in the


The total


number of Japanese immigrants to Brazil during the 33 years from 1908 to


1941


was estimated at 190,000 (Fujii and Smith,


1959).


The Period 1952-1958









Brazil (due to Japan's


involvement in the war) and also because Brazil


adopted a quota system to restrict all foreign immigrants.


After 1952


, Japanese


immigration to Brazil resumed,


1960s.


although at a much lower rate, until the late


It is worth noting that during the four years from 1953 to 1956,


Japanese immigration sped up rapidly and the Amazon region received a
larger proportion of the total of approximately 14,000 immigrants. The

annual proportion of Japanese immigrants of all the immigrants rose steadily


from 2.4


in 1953 to a postwar high of 11.0% in 1956.


In 1958, an important census was conducted by a special commission of
Japanese immigrants with financial support from the Japanese colony in

Brazil, the Brazilian government, the Japanese government, the Population

Council of New York and various private enterprises. In commemoration of
the fiftieth anniversary of Japanese immigration to Brazil, the census

provided valuable information on the Japanese immigrants and their


descendants.


The census organizers planned to cover "information not only


on the present situation of immigrants and their descendants,


but on the


immigrants'


background in Japan,


their initial conditions in Brazil, and the


changes they had undergone in the 50 year period" (Suzuki,


1965:117).


project was, in fact, a monumental work on various aspects of Japanese
immigrants and their descendants in Brazil.

According to the 1958 Japanese self-census, there were a total of 429,413


Japanese, of whom 32.3


were immigrants and 67.


were their


descendants.


Meanwhile


44.9


of the Japanese resided in urban areas and


55.1% lived in rural areas.


Proportionally


slightly fewer immigrants (42.9


*









The Period from 1959 to the Late 1960s


No statistics are available on the number of Japanese immigrants to

Brazil during the period from 1958 to the late 1960s, when large-scale


immigration from Japan to Brazil virtually stopped.


Nor is there any


consensus among researchers on the actual number of immigrants for this


period.


Sims (1972) reported one interesting feature of the Japanese migration


to Brazil during this period:


the Brazilian government authorized two


Japanese-Brazilians to import immigrants from Japan to certain areas in


Brazil and set them certain quotas as well.


For example,


"the late Mr.


Yasutaro Matsubara was authorized to settle 4,000 Japanese families in central


Brazil (southern Brazil was approved later) and Mr. Kotaro


Tsuji was


authorized to settle 5,000 Japanese families in the Amazon region" (Sims,


1972:246).


These quotas remained effective until 1966, when the "Japanese-


Brazilian Joint Committee" was established and the quotas were abolished.
Japanese agencies, governmental and private, continued to provide subsidies


to immigrants, especially those bound for Brazil during this period.


Suzuki


(1981) estimated the total influx for the period from 1952 to the late 1960s at

50,000, while Smith (1979) estimated it at 60,000.


All tolled


, during the 50 years from 1908 to 1958, about 240,000 Japanese


migrated to Brazil and the majority of them settled in the state of Sao Paulo


(Fujii and Smith,

the total amarelo


1959; Suzuki,


1981).


population was 329,082,


1950 Brazilian census reported that


and 84


of them resided in the


state of Sao Paulo.


The 1958 census of the Japanese community "reported that









Social Characteristics and Social Mobility of the Japanese Immigrants


Japanese immigrants were brought to Brazil primarily as farm laborers.
As a result, the majority of them were at the bottom of the social hierarchy


when they started their new lives in the new country.


Here I will focus on


the initial social status, as marked primarily by their occupations, of the
Japanese immigrants and the changes in the distributions of industries and


occupations for them during their first fifty years in Brazil.


Then I will


review their initial educational status and how that changed through the


years.


I wil


also review some demographic characteristics of the Japanese


immigrants that are closely associated with their social mobility.


Occupational Distribution and Mobility


The occupational distribution of immigrants to Brazil first and
foremost reflected the Brazilian immigration policy at the time, i.e., the
creation of a small-farming class and the provision of a supply of cheap labor

for plantation owners.


During the prewar period from 1908 to 1941,


98.8% of the Japanese


immigrants to Brazil were classified as farmers, whereas only 59.6% of all


immigrants to Brazil were farmers.


The proportions of farmers among the


arger immigrant groups are 78.6% Spaniards,


49.0


Italians and 47


Portuguese (Fujii and Smith,


1959).


During the postwar period from 1954 to


1956


, the percentage of farmers among Japanese immigrants dropped to about









Suzuki (1981) reported that 94% of all family heads started as farmers,
of whom were at the lowest status as colonos primarily on coffee


plantations (90%).


However, in 1958, the proportion of farmers among the


Japanese immigrants dropped to about 61


dropped to only


of the total farmers.


% and the proportion of colonos
The majority of the former colonos


went to large urban centers to work as craftsmen and unskilled laborers.
The 1950 Brazilian Census provided the first systematic information on


the distribution of industry by racial group.


Smith (1972) included a very


detailed table on the distribution of industry for males 10 years of age and


over by color, based on the 1950 census data.


Let me briefly summarize the


industry distribution of the amarelos and the standing of this group relative
to the other races described in Smith (1972).

The 1950 census included eleven categories of industries, but the
distribution of industry for the amarelos was highly concentrated in the
following four categories: agriculture (which included forestry and fishing)


(69.0


service (10.2


wholesale/retail trade (10.1


(which included construction and processing) (6.0%).


) and manufacturing
The proportions for the


remaining industries were, in descending order: transportation (which


included communication and storage) (2.3


finance (which included


insurance and real estate) (0.9%),


industries (0.


liberal professions (0.6%),


and social activities (0.4).


extractive


The total number of people who


were engaged in


"public administration, legislation and justice" and


"national defense and public security" was so small (90 and 138 respectively)


that they were omitted in the percentages for the original tabulation.











Table


Industry Distribution of Brazilian Males Aged 10 and Over by Color, 1950


Industry


Total (%)


White


Negroes


Yellow


Pardos


Agriculture
Extractive


64.6


70.0


69.9


Manufacturing
Wholesale


Finance


Service
Transportation
Liberal Profession
Social Activities
Public Ad.
National Defense


0.04


0.09


All Industries


100.0


100.0


100.0


100.0


100.0


Source: Table XI in Smith


, 1972, pp. 94-95.


We can also look at the proportions of employers, employees, self-


employed workers,


and family workers by race and see the differences among


racial groups.


Table


illustrates the proportions of different employment


statuses by color for all industries and agriculture, the most important


industry in 1950.


In both all industries and agriculture, the amarelos,


compared to the other groups, have the highest proportions of employers


(11.8


and 10.8% respectively) and the highest proportions of family workers


(29.6


and 38


respectively).


Expectedly, they have the lowest proportion


of employees (23.


in all industries and 15.4


in agriculture) among the









Table


Employment Status of Brazilian Males Aged 10 and Over
for All Industries and Agriculture by Color, 1950


Industry


Total(


White


Negroes


Yellow


Pardos


Industries


Employers
Employees
Own Account


46.3


23.7


36.3


Workers


Family
Total


32.0


Workers


100.0


Agriculture
Employers
Employees
Own Account


3.4
34.5


31.0
16.9
100.0


4.5
31.9


13.5
100.0


34.9
29.6
100.0

10.8


48.3


18.5
100.0

2.0
34.2


Workers


Family
Total


Workers


37.4
24.7
100.0


37.4


100.0


32.0


100.0


35.3
38.5


100.0


39.8
24.0
100.0


Source:


Table XI in Smith


, 1972, Pp. 94-95.


1958 Japanese self-census provided valuable information on many


aspects of their lives as a social group.


Tables


2.3 and 2.4 are calculated and


abbreviated from Table


7 in Suzuki (1965) to give more focused analysis on


the occupational distribution of the Japanese immigrants and their


descendants in 1958.


Table


3 shows that the proportions of farmers among


men and women in the labor force for the total population are approximately


the same


for men and


57.7%


for women.


Nevertheless


there are


noticeable differences between the immigrants and descendants and also


between the two


sexes


for the immigrants.


The proportion of farmers for the


male immigrants is 60.3


, while that for the male descendants is only 54.0%.


The difference in the proportion of farmers by


sex for the immigrants is










Table 2.3
Proportion of Farmers among Japanese Immigrants and Descendants


Aged 10 and Over in Labor Force by Sex,


Brazil


1958


Immigrant
Status


Males


Total


Farmers(


NF*(


Total


Females
Farmers(


NF*(%)


7,893


42.4


33,224


57.7


42.3


Immigrants

Descendants


67,518

50,375


60.3

54.0


39.7

46.0


11,492

21,732


66.6

53.0


33.4

47.0


Source:


*NF


Table


7 in Suzuki


, 1965


Population Index, 31:2,
= nonfarmers


, "Japanese Immigrants in Brazil,"
p.135.


Note:


There were three categories,


"farmers"


nonfarmerss"


and "farmers


and nonfarmers," in the original table.


For convenience and clarity,


the first and third categories are combined into "farmer" here, and the
second category remains the same.


Table 2.4 indicates the overall occupational distribution,


including the


most important one, farmer, for the population as a whole and for


immigrants and descendants separately.


Since the category of farmer here


excludes those farmers who had nonfarming jobs, not like the one used in


Table 2.3


, the percentages of farmers for all three groups are consistently a


little bit lower than those in


(less than 1


Table


However, the variations are minimal


) and the basic pattern remains the same.


The exact


percentages of farmers for the total population, immigrants,


and descendants


are 56.0


58.6


, respectively.


flnr i-he frk4til nnnIII n it r-tnL. innc iin-


-In rw n f Ct 4-rplr ^t nn^iy'^i/r


nW~lltl; lfl- nn ll/









remaining 0.4


under the category of "other" belongs to occupations


classified as


"fishermen."


"miners," "quarrymen"


"unqualified


laborers"


in the census.


For the immigrant group, the order remains the same for all


the occupations, except that the order for "clerical" and


"transportation / communication"


is reversed:


1) salesmen (17


craftsmen (10.4


3) service (5


.2%),


transportation/communication (2.0


4) professional (4.2

0) and 6) clerical (1


occupations within the category of "other"


account for 0.6


The three


of the


immigrants.
There are some interesting changes in the occupational distribution for


the descendant group.


First, the percentages for both salesmen and craftsmen


rank first and are identical to one another.


Second, the proportion of clerical


workers exceeds that of professionals, with the others more or less in the


same order as those for the other two groups.
proportions for the occupations are as follows:


More specifically,


salesmen and craftsmen


(14.4%), service (5.4%),


clerical (5.2%)


, professionals (4.6


) and


transportation/communication (2.4%).


The remaining 0.4% is distributed


among the three occupations mentioned above.
The overall trend in the changes of occupational distribution from the


immigrant to descendant group can be summarized as follows:


1) The


proportion of farmers and salesmen decreased from the immigrant to

descendant group; 2) there were large increases in the proportions of

craftsmen and clerical workers among the descendants, and 3) there was a

slight increase in the proportions of transportation/communication,









Table 2.4
Occupational Distribution of Japanese Immigrants and Descendants


Aged 10 and Over in the Labor Force, Brazil,


1958


Occupation


All (%)


Immigrant Status
Immigrants (%)


Descendants (%)


Farmer


56.0


58.6


53.2


Professional/
Technical
Clerical


Sales


Transportation
Crafts


Service
Other


Number


150,170


78,585


71,585


Source:


Table


7 in Suzuki


1972


, "Japanese Immigrants in Brazil,"


Population Index,


Note:


31:2, p.135.


There were ten occupations in the original table, in addition to the


seven listed here. Due to the


space limit and the insignificance of the


three categories


"fishermen," "miners, quarrymen" and


"unqualified


laborers," they are combined under the category of "other" in this table.


Suzuki (1981) described the change in the employment status of

Japanese immigrants and their descendants by classifying them into two


broad categories: independent persons and employed persons.


For farmers,


colonos and sharecroppers were considered employed persons and tenant

farmers and land-owning farmers were regarded as independent persons. For

nonfarmers, employed persons included employees and independent persons


included emnlovers and the


self-emnloved.


Suzuki wrote.


"Whereas the









The distribution of industries in Table

picture from a slightly different perspective.

original table from Suzuki (1965), but I have


2.5 provides us with a similar

There were ten categories in the

listed here only six of them,


which


, by the way


cover almost 99.0


of all industries.


Needless to say,


agriculture has the highest proportion of workers for all three groups:


for all


,59.8% for immigrants and 54.3


for descendants.


Apart from


agriculture,


for both the immigrant and descendant groups, the order of the


industries with the highest to lowest proportion of workers is the same.

Therefore, the order of industries for the total population is the same as well.


They are, in descending order, trade (17.5%),


service (13.3


manufacturing


transportation (2.4%) and finance, insurance, real estate (1.4


remaining 1.0% under the category of "other" belongs to the


industries of


"government," fisheri


es" and "mining."


They are not listed here because


they are negligible in terms of percentage.


Table


Japanese Immigrants and Descendants Aged 10 and Over
in the Labor Force by Industry, Brazil, 1958


All (


Agriculture
Manufacturing
Trade
Finance
Transportation
Service


Immigrant Status
) Immigrants (%)
59.8


Descendants (


54.3


Other


Number


Source


Tahlo


150,170


7 inSzk


*A A L t&L.'At *t A .tLALC LJ I UJ J


78,585


. "Taanese Immigrants in Brazil."


Industry


:


I









Although the proportions of industries have exactly the same ordering

for both the immigrant and descendant groups, there are variations in the

exact proportions of all industries for the two groups. Apart from a decrease

in the proportions of farmers, the descendants have increases of various

degrees in the proportions of workers for all the industries except that of


trade


which has a loss of 1


(18.3


for immigrants to 16.6


descendants).


The two biggest increases of workers occur in the industries of


service and manufacturing for this group; the former increases by 4


(from


11.3% for immigrants to 15.


for descendants) and the latter by 1


(from


for immigrants to 8.0


for descendants).


The agricultural status of postwar migrants and their descendants in


1958 is described in Sims (1972).


The study, based on a survey of 4,268


Japanese farmers who arrived in Brazil during the period 1952-58, showed


that of the total sample, 51


were colonos,


16.8% had become owner-


farmers, 15.4


had become renters, 16% had been reduced to sharecroppers


and 0.6


had become farm administrators.


In comparing the prewar and


postwar migrants in terms of ownership of land, Sims noted that "the private

ownership of land was slightly more common among the prewar migrants


(22%) than their postwar successors (16.8%)"


(1972:250).


However, one


important fact about the prewar Japanese migrants was that only 1.3


them were still colonos by


1958.


Another significant characteristic of the


Japanese farming community in the postwar period was that family workers


took up 59.3


of all the farmers


, "revealing the dependence of the farm


families upon their own kin" (Sims,


1972:250).









By the late 1950s, among nonfarmers, craftsmen constituted less than a


quarter, the proportion of salesmen increased from 8


service workers accounted for 10


to more than 50


Unskilled laborers used to account for


almost a quarter of the total nonfarming Japanese population, but by the late

1950s they had virtually disappeared (Suzuki 1981).

The occupational status of nonfarming Japanese Brazilians was

described in Sims (1972), who compared the prewar and postwar groups (see


Table 2.6).


There were striking differences between the prewar and postwar


migrants in terms of occupational status; 81


working for themselves, i.e.,


of the prewar migrants were


they were either employers or self-employed,


while only 28.5% of the postwar group were doing so.


By the same token,


nearly two thirds of the postwar migrants were employees or working for


others, whereas only


16.4% of the prewar group were so.


The only advantage


of the postwar group over the other was their higher proportion of managers


vs. 2.


) due to an increased level of education and more diverse


backgrounds among the postwar immigrants.


Table


A Comparison of Occupational Status
of Prewar and Postwar NonFarming Japanese Immigrants


Occupational
Status
Self-employed
Employers
Employees
Managers


Prewar Immigrants


Postwar Immigrants


59.8


63.9


Source: Sims (1972)









in Brazil.


The most obvious change was the sharp decrease in the proportion


of farmers, from over 95


in the beginning decades of immigration to about


56.0% in the late fifties.


Second


, the percentage of colonos in the prewar


Japanese migrants decreased from about 80


for the period before 1941 to


in 1958.


Third


, the amarelos,


who were overwhelmingly made up of


people of Japanese origin,


exceeded all other racial groups in the proportion of


employers.


Fourth


, they maintained the tradition of working as family units,


which had advantages over individual workers in terms of utilizing human
and capital resources.


Diversification of Agricultural Crops and High Productivity


Japanese-Brazilians are also considered to be "the first to move toward


the diversification of crops in the Slo Paulo coffee-lands"


1990:187).


(Dwyer and Lovell,


In addition to coffee, the Japanese owner-farmers produced cotton,


rice, potatoes and other new crops.


A survey of 35,871 Japanese Brazilian


farm families in 1958 revealed that the largest number of Japanese farmers


grew coffee: 17.6% in Sao Paulo and 27.5


in the nation as a whole.


Vegetable


was the second largest crop, with a farming population of 13%.


third with a farming population of


Cotton ranked


The majority of both the vegetable


growers and cotton growers were in Sao Paulo (Sims,


1972).


The above study also described the employment status of the Japanese-


Brazilian farmers in


1958.


Seventy-five percent of the coffee growers,


43.9


the cotton growers,


about 50% of the poultry raisers and nearly one-third of









It was not only the high proportions of the Japanese farmers in the
above agricultural sectors that is important, but also their production that is

more important in terms of their contribution to the agricultural


development of Brazil.


By the late


1930s, the Japanese-Brazilians accounted


for more than 50% of the cotton produced in the state of Sao Paulo (James,


1937, cited in Dwyer and Lovell,


1990) and were responsible for 80


of the


vegetable production in the suburban area of Sao Paulo city (Makabe,


In 1958, Japanese-Brazilians,


1981).


with only one percent of the total farming


population,


produced about 62


of the tomatoes


of the peanuts,


27%


the potatoes, about 12


of the eggs, and about 12


of the cotton produced in


Brazil.


They were also responsible for about 93% of the tomatoes, 92% of the


tea, 68% of the potatoes, 43


of the peanuts, 37% of the eggs, 36% of the


peppermint,


27%


of the cotton


of the banana produced in the state of Sao


Paulo (see


Table 2.7).


Table


Agricultural Production of Japanese-Brazilians
in SAo Paulo and Brazil by Crop, 1958


Crop % of the Brazilian Total % of the Sao Paulo Total

Tomatoes 61.7 93.3
Peanuts 39.1 42.8
Potatoes 27.0 67.9
Eggs 11.6 37.0
Cotton 11.6 26.8
Coffee 5.9 7.1
Banana 5.3 21.8
Fruits 2.9 --.
Rice 2.3 8.1
rT- n 1








Studies on the social mobility of the Japanese Brazilians in the last two


decades are extremely rare in the English language publications.


study available is Dwyer and Lovell (1990),


One such


"Earning Differentials Between


Whites and Japanese:


The Case of Brazil"


This study uses a sample of 272


white males and 242 Japanese males ages 18-64,


1980 census of Brazil.


from the 0.8% sample of the


The main findings of this study are: (1) the average


earnings of Japanese males are 61% higher than that of whites; (2) 48% of
Japanese males have more than nine years of schooling compared to 24%


white males; (3) only 51


of the Japanese are workers whereas


of the


whites are workers; (4) three times as many Japanese as whites are employers


and 32


of Japanese versus 19% of whites are self employed.


These findings


suggest that Japanese-Brazilians have surpassed whites in terms of many
important social indicators.


Educational Status


Educational status of a population is usually measured by its literacy
(illiteracy) level and the percentages of people who have received elementary,


secondary and higher education among its literate people.


There is ample


evidence that from the very beginning, Japanese-Brazilians fared very well in
terms of educational status among the various immigrant groups and among
the racial groups as well.
The literacy rate of the Japanese immigrants was one of the highest
among all immigrant groups through time. For the prewar period 1908-1941,









literacy rate was then measured in terms of the ability to read and write in the


native languages of immigrants,


not in Portuguese, the official language of


Brazil.

However, according to the 1940 and 1950 census, the illiteracy rate for

the yellow people was the lowest among the four racial groups (Smith,

1972:490):


Race


1940


1950


Yellow
White
Pardos
Negroes


34%
47%
71%
79%


It is also worthwhile to point out that the illiteracy rate declined by 50%

among the yellow people, while its rate of decline was not as great among the

other three groups, especially among pardos and Negroes.
The 1958 Japanese Self-Census indicated that of all Japanese residing in


Brazil aged


7 and over, the illiteracy rate was only


2.5%;


in urban areas


and 3.8


in rural areas.


The census also provided information on this subject


for immigrants and descendants separately: the illiteracy rate for all


immigrants was 1


with 1


in urban areas and nearly


in rural


areas.


Interestingly, the illiteracy rate for descendants was slightly higher than


that for immigrants: 3.2%

4.1% in rural areas (Suzuk

Sims (1972) reported


for all descendants, with


in urban areas and


1972).

he result of a 1962 survey of 151,701 newspaper


readers over 14 years of age to show the literacy rates in both Portuguese and


I









computation, that "at least


.4% of the community surveyed read


Portuguese, while a minimum of


were literate in Japanese in


1962"


(1972:258).

According to the 1950 census, the proportions of people who completed


elementary schooling was much higher among the


Yellow population than


was the case nationwide (


vs. 17.9


) (Smith and Fujii, 1959).


proportions of people who attended different levels of schooling among the

Japanese immigrants and their descendants in 1958 were described in Suzuki

(1965).


Table


2.8 offers a


ummary of the above information:


In urban areas


67.3


of the people aged


7 and over attended primary school,


29.2%


attended


secondary school, and 0.


attended college, while in rural areas, the


corresponding figures were 82.6%,


11.8% and 0.8%.


When immigrants and


descendants were compared,


areas


example,


the latter did better than the former in urban


whereas the former did better than the latter in rural areas.


the percentages for primary and secondary schooling among the


urban immigrants were


.3 and


counterparts were 62.6 and 33.9.


21.0, while the same percentages for their

On the other hand, proportionately, more


rural immigrants attended secondary school (14.


counterparts (9.9).


) than did their


The proportions of people who attended college for all


groups was less one percentage.
Suzuki (1981:65) noted that "a relatively high educational level in

comparison to that of the society on a whole would seem to lessen handicaps


affecting foreign immigrants in their struggle for a better life."


The high









Table 2.8


Japanese Immigrants and Descendants Aged


7 and Over


by Level of Education and Residence, 1958


Residence


Total


Primary


Secondary


Higher Ed.


Urban


Immigrants
Descendants

Rural


Immigrants
Descendants


160,796
58,972
101,824

189,565
77,610
111,955


67.3
75.3
62.6

82.6
81.2
83.7


29.2
21.0
33.9

11.8


Source: Suzuki, 1965


Demographic Characteristics


The most distinctive demographic feature of Japanese immigration to


Brazil was "family immigration," which

imposed by the Brazilian government.


h was the direct result of a regulation

According to this regulation, an


immigrant family must have at least three capable laborers who were above


fifteen years of age.


Consequently, about 95


of the Japanese immigrants


between 1908-1941 and 80',

groups, as compared to 64


between 1954-1956 came to Brazil in such family

and 54% of the total immigrant population in


these two time periods (Fujii and Smith,


1959).


As a correlate of the high proportion of family units among the

Japanese immigrants, the percentage of married people was also high among


4Knnwv eit-I1 l4- nrA~n Anrrns~nn ,.y44, 4-le^n Cd aC i-I, n nrandn C., 4


nC{ C -n / nmrf^'^ffcn* C^^y ^4 ^nn^~


*









33.5% and 2.1


The proportions of singles and widowed increased by 8.4%


and 0.4, whereas the proportion of married decreased by 8.8%.


This was


probably resulted from the relaxation of family unit rule applied to Japanese

immigrants during the late fifties.

Table 2.9 illustrates the marital status of the Japanese population aged


15 and over by sex and generation in 1958.


of the males and 35.9


For the whole population, 44.5%


of the females were single, while 52.3% of the males


and 56.6% of the females were married.


The proportion of married people


was up by more than ten percent from 42.3% in the period 1908-1941


. The


percentage of married people among the immigrants was even higher due to

the fact that most of the immigrants were adults and had become parents or


grandparents by 1958.


The proportion of married people for the total


population was heavily affected by that of the immigrants since they were still


the majority at that time.


In contrast, the percentage of married people


among the second generation of Japanese was far lower than that for the


immigrants,


due to their relatively young age.


Table 2.9
Marital Status of the Japanese Population in Brazil
by Sex and Generation, 1958


Immigrant
Status


Males


Sin.


Mar.


Sep.


Females


Wid.


Sin.


Mar.


Sep.


Wid.


All Japanese
Immigrants


44.5


35.9


79.4


56.6
79.9


Generation


35.2


3rd & 4th









However, one common element among all groups, immigrants and

descendants alike, was that proportionately more women were married than


men; 56.6


vs. 52


for all Japanese, 79.9


21.4% for the second generation.


% vs. 79.4 for immigrants, and
This was largely caused by the fact


that women married at younger ages than men did in general


. Therefore, the


sexual differences in the percentage of married people among different groups

was a main indicator of the mean age at marriage for the groups concerned.

Family type and structure are known to correlate with the


socioeconomic well-being of a particular group.


Suzuki (1981) showed a


positive correlation between the Japanese family structure and the
improvement of their economic status by comparing the frequency of family

types with their economic status expressed in terms of the employment status


of the family head and property ownership (see Table


2.10).


He found out that


among the land-owning farmers, 36


were "lineal"


were three-generation families and


and "lineal and collateral" families; among the tenant


farmers, the corresponding figures were 21


and 24%.


In contrast, the


percentages of three-generation families and "lineal" and "lineal and


collateral" families among the sharecroppers and colonos were


respectively.


16%


Therefore, we can conclude that more independent


farmers tend to have extended (three-generation) families and lineal or
lineal/collateral families than employed farmers (sharecroppers and colonos).

This, in turn, suggests that three-generation families, and lineal and

lineal/collateral families may have a positive effect on the employment


status, i.e.


, whether being an independent or employed person.









and employers (13%,


This suggests that larger families may not be an


advantage for non-agricultural workers.


On the other hand


there was a


positive association between the value of property owned in both rural and

urban areas with three-generation families and lineal and lineal/collateral


families.


In other words, the proportion of three-generation families and


lineal and lineal/collateral families increased with the increase in value of


property owned.


According to Suzuki (1981),


in rural areas, the proportions


of three-generation families for those who owned no property


low property,


medium property and high property were 18%,


28%


42% and 53%


respectively, whereas for non-farmers in urban areas, those proportions were


24%


37%


respectively.


The same pattern remained for lineal


and lineal/collateral families (see


Table


Table 2.10
Proportion of Traditional Families among Japanese Heads of Family


by Employment Status for Farmers and Non-Farmers in Brazil,


1958


Employment Status Three-Generation Lineal and Lineal/Collateral
Families (%) Families (%)
Farmers
Landowners 36 40
Tenant Farmers 21 24
Sharecroppers 16 20
Colonos 10 11
Non-Farmers
Employees 29 31
Self-Employed 23 27
Employers 13 17


Source: Suzuki


1981









cooperation among family members.


Such cooperation is effected,


inter alia,


through family labor, i.e.,


family members work without wages in an


establishment operated by the head or another family member"(1981:69).


Table


Proportion of Traditional Families among Japanese Farmers and


Non-Farmers in Brazil by Value of Property Owned,


1958


Value of Property Three-Generation Lineal and Lineal/Collateral
Owned Families (%) Families
Farmers
None 18 21
Low 28 32
Medium 42 46
High 53 57
Non-Farmers
None 17 26
Low 24 27
Medium 35 36
High 37 37


Source: Suzuki


1981


Closely related to the family type and marital status of an immigrant


group is its


ratio, which is even more important when there are relatively


few inter-groups marriages.


During the period 1908-1941,


sex ratio of


128:100 among Japanese immigrants was significantly lower than that of any


other major immigrant groups ( 146 for Spaniards,


175 for Germans, 183 for


Italians


, and 208 for Portuguese).


However


sex ratio of the Japanese


immigrants rose to 1


for the period from 1954 to 1956 due to the relaxation


of the regulation on family groups.









Japanese immigrants were 12 years of age or younger


compared to 23


among the total immigrants,


during the period 1908 to 1941


(Fujii and Smith,


1959)


1950 Brazilian Census indicated some changes in some of the


demographic characteristics


of the amarelo population.


For example,


over


of the amarelo population were under twenty years of age, and the


ratio for them dropped from 128 to 110.8.


The fertility ratio (number of


children under five years of age per 100 women aged 15-49) for amarelos in
1950 was 79.6, the highest among the four major racial groups (65.3 for white,


55.6 for Negro, and 69.2 for brown).


On the other hand


, the proportion of


Asians in the Brazilian population remained at 0.6% from 1940 to


1950


(Smith


, 1972).


This may be due to the pause in Japanese migration to Brazil


during the period 1942 to 1952.

The 1958 Japanese self-census offered information on the changes of


the characteristics of the Japanese population at the time:


sex ratio was


108, a decrease of 2.8 from 110.8 in 1950; the number of people under 15


of age was 40.5% of the total population,


years


indicating that the population


became younger than it was eight years ago; and rural residents accounted for


about 55% and the urban residents 45


from rural areas (Suzuki,


, showing large volumes of exodus


1972).


Summary


Japanese migration to Brazil started at the turn of the century because


sex









during the period of initial industrialization and urbanization in Japan. Al
the same time, Japan faced strong resistance against overseas emigration in


countries like Australia, the United States, Canada and Peru.


By contrast,


Brazil sought after Japanese farm laborers because of a severe labor shortage


on coffee plantations after the abolishment of slavery


in 1888, and


particularly, in 1902 after the Italian government ceased subsidizing the
migration of its agricultural laborers to Brazil.

Japanese immigrants were subsidized by the state of Sao Paulo from
1908 to 1923 and then by various Japanese emigration agencies, both private


and governmental,


up to the late 1960s.


During the period 1908-1941,


approximately 190,000 Japanese immigrants came to Brazil. After a ten-year


pause from 1942-1952 due to World War II,


the migration wave continued at


a much lower rate until it virtually stopped in the late 1960s.


The total


number of immigrants during this period was estimated at 50,000-60,000


(Smith


, 1979; Suzuki,


1981).


The 1958 Japanese self-census indicated that the


Japanese population in Brazil at the time was 429,413, of whom 32.3% were


immigrants and 67.


were their descendants.


The majority of the Japanese immigrants were farmers and started as


colonos on coffee plantations in the state of Sao Paulo.


They rose from the


lowest and least privileged status of colonos to the middle class status in both
rural and urban areas through their hard work during the 50 years after their


first arrival in Brazil.


The experience of the Japanese population during the


1960s and 1970s is proof of their continued success in upward social mobility.
The most cited reasons for the success story of Japanese immigrants are









explanation focuses on the adaptive ability of Japanese, and their traditional


values and characteristics.


It seems to me that the former is mostly related to


external factors, and the latter to internal factors,


from the viewpoint of the


Japanese immigrants.

In comparing the experiences of Japanese immigrants in Canada and

Brazil, Makabe (1981) concludes that the major reason for the success of

Japanese immigrants in Brazil was the lack of economic competition from the

native Brazilians and other immigrant groups and hence the lack of

unfavorable differential treatment in wages because they occupied different


labor market.


He also notes that "ownership of land, which was the highest


achievement to be attained for the immigrants,


became possible relatively


easily and quickly" (1981:800).

In contrast, Dwyer and Lovell (1990) explain the Japanese success

mostly in terms of their adaptive ability, and their cultural values and


characteristics.


They point out three major reasons for their success: (1)


"second generation Japanese-Brazilians quickly learned the language,


business practices, and legal system of Brazil";

placed a great deal of emphasis on education"


(1990:188).


(2) "Japanese immigrants

(3) they were very industrious


This mostly cultural explanation is similar to the one used for the


explanation of the socioeconomic achievement of


(Bell, 1985; Kitano, 1969; Newsweek, 1982


Asian-Americans in the


Petersen, 1971).


These two types of explanation are very important in understanding

the Japanese experience in Brazil, and they are complementary rather than


mutually exclusive.


However


there were also other factors that contributed


,


J


v









industrialization in Sao Paulo, the continued close connections with the

home country and financial and technological assistance from the home

country, the establishment of agricultural cooperatives and ethnic enclaves,
and lack of overt racial discrimination by the Brazilian society,


By human capital, I refer to the educational level,


knowledge of


farming and technological skills the Japanese immigrants and their


descendants possessed.


As shown above, the educational status of the


Japanese population was the highest among the four census racial groups and


they had advanced knowledge of intensive agriculture.


These attributes


translated into better adaptation to the new environment and greater


efficiency and higher productivity


which would certainly result in higher


economic profits.
The Japanese immigrants benefited tremendously from, as well as


contributed to,


"the remarkable economic and demographic growth of SAo


Paulo attributable to the coffee industry and industrialization" (Tsuchida,


1978).


The labor shortage on coffee plantations brought them to Brazil in the


first place, and the urbanization and industrialization in the state of Sao
Paulo offered them the opportunity of pioneering vegetable farming and the


poultry industry.


The development of the textile industry in Sao Paulo in the


early 1930s created a huge domestic market for cotton,


and Japanese Brazilians


dominated the cotton industry from the outset due to their expertise in
growing cotton.

Meanwhile, Japan's importation of large quantities of cotton in the
mid 1930s from Sao Paulo also helped the Japanese Brazilian cotton growers.









already enabled the Japanese to solidify their economic base in such a way that
they and their descendants in Portuguese America could securely stand on
their own feet in total isolation from their mother country" (1978:311).
The fact that the Japanese immigrants maintained close ties with and


received financial and technical support from their home country


especially


in the early years,


was very important to their success in Brazil.


The Japanese


government and private companies financed various colonization projects
and provided information, technical assistance in farming, and even
improved seeds, which greatly promoted land ownership and increased


agricultural productivity among the Japanese immigrants (Tsuchida,


1978).


Another important feature of the Japanese immigrants in Brazil was
that from the outset, they established their own ethnic enclaves in the form


of agricultural cooperatives and larger community settlements.


Normano


and Gerbi described the Japanese in the following way:

The Japanese live almost completely isolated from the native


element in Brazil.


The population of their centers varies from


three hundred to six or seven thousand


in cities


, towns, and


large fazendas, but always they remain in atmosphere and
surroundings completely Japanese (1943:39).


Their agricultural cooperatives facilitated the transportation and the
marketing of their products, and the ethnically homogeneous communities


provided them


"with adequate educational opportunities, medical care,


technical assistance


loans


, and above all, a sense of security" (Tsuchida,


1978:313).


There is no doubt that these ethnic associations played a major role









racial discrimination by the dominant society as were their counterparts in


North America (Daniels,


1977


Daniels, 1988; Kitano and Daniels, 1988; Lee,


1989).


At least, there was


no overt discrimination against them in the


economic sphere so that they were able to demonstrate fully their valuable
assets and compete on an equal footing with others for land ownership,


property and social mobility.


Brazil


On the subject of anti-Japanese sentiment in


Tsuchida wrote:


Devoid of any serious economic conflict between the Japanese
community and the dominant society, charges against this
ethnic minority centered around racial desirability and the
intangible threat of Japanese imperialism. Anti-Japanese
agitation was restricted to a small circle of intellectuals who
advocated Japanese exclusion, on ideological ground, rather than
economic and political reasons (1978:321).


On the other hand


, the Japanese immigrants didn't compete with the natives


for occupations then considered more favorable, such as commerce.

apparently avoided possible conflicts in their economic activities. T


first engaged in coffee growing, then pioneered cotton,


They


They were


vegetable and fruit


farming, all of which were much needed by the Brazilian society


. In other


words, they had their own labor market, and were not in direct competition

with the dominant society.

However, this does not imply that Brazil has been a racial democracy,


as some scholars have advocated.


In fact, there is a body of literature that


indicates the scope of racial inequalities in Brazil (Hasenbalg, 1985; Lovell,


1989; Lovell and Dwyer, 1988


Silva, 1978; Silva,


1985; Wood and de Carvalho,


r









They managed to rise from the bottom of the society and achieve middle class
status within the first fifty years of their arrival mainly by hard work, assets in
human capital, a traditional practice of working as family units, demographic


factors (relatively balanced


sex ratio and younger age structure), collective


efforts and ethnic unity, strong support from the home country, a favorable
economic situation in Brazil and a lack of overt discrimination against them,
especially in the economic and political arenas.













CHAPTER 3
FERTILITY DIFFERENTIALS


AMONG


ASIAN


WHITES


AND


AFRO-BRAZILIANS


A Brief Review of Literature on Fertility Studies


Human fertility behavior is the subject of study in many disciplines of


the social sciences, and various theories on fertility


have been put forward.


Some of the major fields of study that deal with human fertility are


demography, sociology, economics, anthropology


psychology and biology.


Each discipline tends to focus on slightly different aspects of human fertility
behavior and differs somewhat in its approaches due to its distinct theoretical


orientations and scopes of study.


However, there are many things that


fertility studies have in common.
The economic theory of fertility is perhaps the most influential among


competing theories.


The most important works of this school of thought are


Leibenstein (1957),


Becker (1960),


Easterlin (1969) and Schultz (1973).


applying the economic theory of consumer behavior to childbearing


decisions


, they regarded human fertility as a result of rational decision based


on an effort to "maximize satisfaction, given a range of goods,


their pri


and his own tastes and income"


(Easterlin


, 1975:54).


In other words


"children









things being equal, higher income usually results in higher fertility rate; (2)

an increase in the price of children relative to other goods results in lower

fertility.

Counter to the first hypothesis, cross-cultural and cross-sectional
demographic data generally show that higher income groups tend to have


fewer children compared to lower income groups in a country.


Similarly


aggregate data show that more affluent and developed societies tend to have


lower fertility rates than their less developed counterparts.


noted


It should be


, however, that these studies may not represent an adequate test of the


economic theory of fertility (which predicts a positive correlation between


income and fertility).


The reason is that aggregate data on fertility rates by


income classes do not measure what economists refer to as the "pure income


effect."


That is, the effect of income after controlling for contraceptive


knowledge and other determinants of fertility behavior.
The second hypothesis is valid and supported by some historical


demographic data.


Yet it offers little insight to differential fertility among


various sub-populations of a society if we assume that "the price of children
relative to other goods" is, more or less, the same for all the people in the


same region at a certain period of time.


Moreover, the economic theory of


fertility analysis leaves little room for the role of sociocultural factors and

other institutional constraints in the fertility decisions and behaviors of
individuals, who live in a complex social context and are bound to be

influenced by many external factors.

In addition, Easterlin (1975) and Todaro (1981) have applied









includes both subjective and objective costs, as well as the time and money
required to learn about and use specific techniques for limiting fertility.

The sociological theory of fertility is mainly represented by Davis and


Blake (1956)


, Davis (1959)


, Freedman (1962),


and Hawthorne (1970).


In this


approach, observed level of fertility is seen as the outcome of the interaction
among biological processes, societal group factors and individual behavior


(Robinson and Harbison, 1980). Social
considerable attention in this approach,
dynamic than the economic approach.


norms about family size are given
and it is broader in scope and more
In an attempt to bridge the gap


between the economic theory and sociological


proposed a new "general theory"


theory,


Caldwell (1976, 1978)


of fertility, which states that


"fertility


behavior in both pre-transitional and post-transitional societies is
economically rational within the context of socially determined economic
goals and within bounds largely set by biological and psychological factors"


(1978:553).


Recent development of this approach is reflected in the


examination of the socioeconomic and proximate determinants of fertility.


(Easterlin, 1983


Standing, 1983; Menken, 1987)


The psychological approach to fertility focuses on individual-level
processes and places emphasis on psychological variables and measures.

Fishbein (1972) argued that human fertility behavior was determined by


people'


intentions, the normative beliefs regarding fertility


and the personal


attitudes toward the importance of these norms.


approach,


here.


Unlike the sociological


norms affect fertility through personal attitudes and intentions


Other works of this orientation include Jaccard and Davidson (1976),









the other factors, economic and social,


affect fertility through individual


attitudes and intentions.
Anthropologists usually study fertility behavior in terms of the

determinants of social and cultural differences within an evolutionary


framework.


Barlett (1980) identified three approaches to fertility within


anthropology: the ecological approach, the cognitive approach and the


statistical aggregate approach.


Chagnon (1968) and Harris (1974) applied the


ecological approach to explain the practice of female infanticide among the
Yanomamo, and concluded that female infanticide was an effective way of


limiting the overall fertility of the group.


Cognitive anthropologists (e.g.,


Marshall


, 1972a and Quinn,


1975) stressed the importance of individual-level


decision making, and attempted to build models for the decision-making


process.


The third approach, the statistical aggregate approach,


people do, not what people say they do" (Barlett,


1980:168).


"stresses what


More specifically,


in this approach,


"an anthropologist observes behavior, records outcomes,


and then analyzes the patterns in the outcomes to construct a statistical


profile of people who choose different options" (Barlett, 1980:168).


Since most


anthropological studies have dealt with relatively homogeneous societies in
the past, they tended to assume that shared values and traditions and societal


norms govern individuals'


behavior


, which in turn determine their fertility.


In short, most anthropological approaches to fertility tend to focus on cultural
patterns.


However


, there is another approach to fertility in anthropology that


stresses the role of material conditions or factors directly related to material









explain the fertility behavior of preindustrial societies. Handwerker (1986)
criticized the cultural approach to fertility as tautological, and offered a


materialist explanation to fertility transition.


Handwerker argued,


cannot identify specific behavioral patterns and the ideas they presuppose


independent of one another.


'explain'


behavior by reference to those ideas


therefore constitutes a covert tautology" (1986:14).


transition occurs "when personal material


According to him, fertility


well-being is determined less by


personal relationships than by formal education and skill training."
Handwerker further explained:

This transformation occurs when changes in opportunity
structure and the labor market increasingly reward


educationally-acquired skills and perspectives,


for these changes


have the effect of sharply limiting or eliminating the expected
intergenerational income flows both from children, and from
the social relationships created by or through the use of children.
(1986:3)

In terms of the relationship between education and fertility, Handwerker

offered an insightful analysis:


education or literacy itself can have no important effect on


fertility.


The linkage between education and fertility is


contingent on opportunity structure, and will turn on the issue
of how material well-being can best be created and maintained,
and how educationally acquired skills and perspectives fit, or do


not fit, into this process.


(1986:18)


The above approaches not only differ in theoretical orientation,


also in unit of analysis.


Both economic and psychological approaches to









which may be an extended family, a clan, a social class or the society as a


whole.


Even when individuals are the focus of attention,


they are situated


within the sociocultural context and regarded as members of a social group,
rather than as isolated individuals acting on their own.




Fertility Differentials among Ethnic/Racial Groups in Modern States


It is well documented in the literature of demography and ethnic/racial
studies that in multiethnic/racial societies, various ethnic/racial groups


reproduce at different rates.


For example, Rindfuss and Sweet (1977) reported


different fertility rates for whites, blacks,


American Indians, Mexican


Americans, Chinese Americans and Japanese Americans in the United States


for the period 1955-1969.


These ethnic/racial groups in the United States


continued to reproduce at different rates for the 1970s (Bean and Marcum,


1978) and 1980s (1980 census, cited in Farley and Allen,


1989).


Fertility


differentials among ethnic/racial groups in Canada were reported in Halli


(1987) and Halli et al.


(1990),


and ethnic fertility differentials in China have


been documented in the Chinese censuses since


1950.


Racial variations in


fertility rate in Brazil are also reported in the Brazilian censuses since


spite of the differences in ethnic/racial composition,


1950.


social and political


system and economic structure among these countries, one common element


about fertility rate is almost universal,


, fertility rate seems to vary along


ethnic/racial lines, as well as alone economic. educational. religious and









better understanding of the factors responsible for differential fertility rates


among groups.


Furthermore, an examination of these factors reveals, among


other things, the nature of relationships between different social groups, be


they racial,


cultural, or economic, or a combination of the above,


in terms of


access to education, level of employment and income, and ultimately the
level of well-being.
Within the larger theoretical framework of fertility research in general,
studies of differential fertility among various subgroups of a population


(Goldscheider & Uhlenberg, 1969; Sly,


1970; Bean & Wood, 1974; Roberts &


Lee, 1974; Gurak, 1978; Gurak, 1980; Johnson & Nishda, 1980; Bean &


Swicegood,


1985) suggest three approaches.


They are the cultural (or sub-


cultural) approach,


the structural (or social characteristics) approach and the


minority group status approach.
The cultural (or subcultural) approach emphasizes the role of values,


norms and ideology in determining a group's


fertility behavior (Goldscheider


& Uhlenberg, 1969).


In this approach,


one "searches for determinants of


demographic variation in the history and cultural traditions of different


subpopulations" (Frisbie and Bean,


1978:2).


Furthermore,


"even when groups


are similar socially, demographically, and economically, minority group
membership will continue to exert an effect on fertility" (Rindfuss & Sweet,


1977:113).


This approach reflects Schermerhorn'


definition of an ethnic


group:


"A collectivity within a larger society having real or putative common


ancestry, memories of a shared historical past, and a cultural focus on one or
more symbolic elements defined as the epitome of peoplehood"









the higher fertility of Mexican Americans stems from the
persistence of cultural norms and values supporting large
families, such as familism---a constellation of norms and values
giving overriding importance to the collective needs of the
family as opposed to the individual---or adherence to the
pronatalist positions of the Catholic church, including
prescriptions against certain forms of birth control.


The social characteristics (or structural) approach does not deny the


possible validity of the subcultural approach,


but it argues that differences in


social status, such as education, occupation and income, account for most or


all fertility differences among sub-groups.


'structural'


This approach also "implies that


assimilation with respect to education, occupation and income


will lead to the elimination of fertility differences between majority and


minority groups" (Bean & Swicegood 1985:7).


The social characteristics


approach has its grounding in the assimilation theory first put forward by


It draws heavily from the idea of "structural assimilation,"


one of seven dimensions of assimilation that Gordon identified


sometimes referred to as "the assimilationist theory."
approach, fertility differentials are attributed to social,


economic characteristics of various groups.


and is


According to this

demographic and


When these factors are


controlled, differences in fertility should disappear.


The minority group status approach was first proposed by Goldscheider
and Uhlenberg (1969), and was thereafter tested and applied in various

studies, such as Sly (1970), Roberts and Lee (1974), Johnson and Nishda (1980)

and Bean and Swicegood (1985). The basic assumption of this approach is that
-- -m -i -


Gordon (1964).









insecurity that accompanies minority group status.


Those minority members


who are in higher socioeconomic standing tend to aspire to greater social


mobility and therefore feel greater insecurity and marginality.


In order to


overcome the feeling of insecurity and the potential obstacles to greater

success, these members are likely to lower their fertility to secure their already


achieved status.


Goldscheider and Uhlenberg (1969) used this approach to


explain the lower fertility rate of highly educated black women as compared


to similar white women.


More recently, Halli (1987) applied this approach to


the fertility of Asian groups in Canada.

Although these approaches differ in focus and have different
theoretical orientations, in my opinion, they actually complement rather


than contradict each other.


They all contribute to the explanation of the


complex causes of differential fertility among various racial/ethnic and/or


socioeconomic groups.


However, it is crucial to test these approaches against


empirical data to determine the most important factors) by examining the
associations between fertility rate and the possible biological, sociocultural


and economic factors.


Specifically, it is important to determine the degrees to


which major independent variables contribute to the fertility level of a


population as a way to


assess


the validity of the competing theories of human


reproduction.


Fertility Differentials Among Asians,


Whites and


Afro-Brazilians









are the main causes for the differences?


The hypothesis tested here is that


socioeconomic status (defined by income level and educational attainment),
rather than color, is the best predicator (but not the only) for differential


fertility among different social groups.


Thus, when income and education are


controlled, color will contribute relatively little to subgroup differences in


fertility level.


It is also assumed that household income and mother's


educational level is negatively correlated with women's


fertility level; i.e.,


higher the income and educational levels are, the lower the fertility level is.


However


, I do not assume that socioeconomic status alone accounts for all


the differences in fertility of various groups.


Therefore, I expect that even


after controlling for the socioeconomic differences, some differences will
remain in the fertility levels of different color groups, although the amount

of variance in fertility explained by ethnic status will be relatively small.

The data set used here consists of women 15-49 years of age only since


we are only concerned with fertility level.


The dependent variable is fertility


level


, and the independent variables are place of residence,


color, age,


education, and mean income.


Fertility level is here defined by the mean


number of children ever born to women of a cohort classified by either color,


age, educational level or income level.


Following the conventional method,


women are divided into either seven age groups (15-19, 20-24, 25-29,


39, 40-45 and 45-49) or four age groups (1


30-34


, 20-29, 30-39, and 40-49) for


descriptive analysis.


In what follows


, I describe the characteristics of the sample data,


compare the mean fertility level by age group, color, educational level,









multivariate regression analyses to examine the relationships among the


variables.


The main findings are summarized at the end of the chapter.


Table 3.1 shows the mean number of children ever born to women in


seven age groups and the standard deviations from the means.

the proportions of each age group relative to the whole sample.


number of children ever born for the total sample is 1.89,


It also shows

The mean


with the expected


increase from the lower to higher age groups.


Table 3.1
Mean Children Ever Born to Women of 15-49 Years of Age
by Age Group, Metropolitan Sao Paulo, Brazil (1980)


A.e Group


Mean


Std Dev


Cases


0.77


20-24
25-29
30-34
35-39
40-44
45-49


3.28


4.17


1.56
1.98
2.46
2.91
3.20


39,916
38,968
33,482
26,925
21,916
19,163
16,283


20.3
19.8
17.0
13.7
11.1


Total


196,654


100.0


Source: Weighted 3% sample data of Metropolitan Sao Paulo, 1980 Brazilian
Census.


The mean number of children by color group is shown in


Afro-Brazilian women have the highest mean (2.18),


(1.82) and Asians third (
statistically significant.


Table 3.2.


with whites second


1.44). Given the sample size, these differences are
The color composition of the women in the sample









Table 3.


Mean Children Ever Born to Women of 15-49 Years of Age
by Color Group, Metropolitan Sao Paulo, Brazil (1980)


Color Group


Mean


Std Dev


Cases


White


147,786
44,365


Afro-Brazilian


Asian


75.3
22.6


4,045


Total


196,195


100.0


Source: Weighted 3% sample data of Metropolitan Sio Paulo, 1980 Brazilian
Census.


Table 3.3 illustrates the mean number of children ever born to women


by age group and color.


Here I still use five-year intervals for age groups in


order to obtain a more detailed picture of the fertility behaviors of the three


color groups.


At every age level,


Asian women have the lowest mean


number of children


, Afro-Brazilian women have the highest mean number


of children, and the mean number of children for white women is above that


of Asians but below that of Afro-Brazilians.


Expectedly, the age group of 1


for all three color groups has very few children,


particularly


Asian


, who, on


average, have only 0.008 children.


Furthermore, the mean number of


children for Asian women ages 20-24 and


25-29


are extremely low; only 0.18


and 0.69 respectively.


In contrast, the fertility levels of whites and Afro-


Brazilians are much higher than that of Asians in these two age groups.
fertility differences among the color groups decrease for older age groups,


the basic Pattern still remain.


In sum. Asian women not only have fpwpr









Table 3.3
Mean Children Ever Born to Women of 15-49 Years of Age
by Age and Color Groups, Metropolitan Sao Paulo, Brazil (1980)


Age Group


Mean


Asian


White


Afro-Brazilian


15-19
20-24
25-29
30-34
35-39
40-44
45-49


0.01*


0.77


3.28


4.17


0.18
0.69
1.56
2.35
2.89
3.51


0.73
1.59
2.39
3.11
3.62
3.93


0.15
0.95
1.95
2.96
4.01
4.81
5.30


Total


Source: Weighted 3


sample data of Metropolitan Sao Paulo,


Census.
*The actual value is 0.008.


In Table 3.4


1980 Brazilian


we see the fertility differences among the three color


groups, controlling for both educational level and age.


There are two


interesting observations to make here with regard to the fertility level of the


three color groups by educational level.


First, fertility differences among the


color groups for women with no schooling are very small (3.63 for Asians,


3.84 for whites and 3.96 for Afro-Brazilians).


Second


, the fertility levels of


Afro-Brazilians at all educational levels, except for the one of no schooling,

are the lowest among the three color groups.

It may seem surprising for Afro-Brazilians to have lower fertility levels

than those of whites and Asians at all educational levels but the first (no


schooling).


This suggests that education may have greater negative imPact on









the disproportionate distribution of Afro-Brazilians in educational level.


Because over 20


of Afro-Brazilians have no schooling, compared to 9.9


whites and 3


of Asians


, their overall fertility level is still higher than


those of whites and Asians, despite their lower fertility levels at all the other
levels.


When the three color groups are compared by age group within the


same educational level


, Asians have children at older ages than do whites at


all levels, and whites have children at older ages than do Afro-Brazilians at


levels of less than 9 years of schooling.


For example, Asian women between


ages 15 and 19 rarely have children at all educational levels, while the mean
number of children for white and Afro-Brazilian women ages 15-19 with less


than


5 years of schooling is more than 0.20.


Furthermore, the mean number


of children for Asian women between ages 20 and 29 ranges from 0.22 to 0.96,
while the mean for white women of the same age group ranges from 0.36 to

2.07, and that for Afro-Brazilian women of the same age group ranges from


0.17


to 2.


At higher educational levels (9 or more years of schooling),


however, Afro-Brazilian women have fewer children than do white women
in all age groups and Asian women in most age groups (see Table 3.4).

The fertility levels of the three color groups, controlling for income


and age, are shown in Table 3.5.


First


between income and fertility level,


fertility levels.


we see the negative association

, lower income groups have higher


The mean number of children for women from the lowest to


the highest income level are


,1.19, and 1.16, respectively.


difference between the fertility level of women in the first and second income










Table 3.4
Mean Children Ever Born to Women of 15-49 Years of Age
by Education, Age and Color Groups Metropolitan Sio Paulo, Brazil (1980)


Years of School


Total


Asian


White


Afro-Brazilian


Zero Years
15-19
20-29
30-39
40-49


Years
15-19
20-29
30-39
40-49


Years
15-19
20-29
30-39
40-49


3.89
0.29
2.09
4.22
5.55


2.30
0.21
1.55
3.02
3.83


0.92
0.08
0.99
2.16
2.70


1 Years
15-19
20-29
30-39
40-49


- Years
15-19
20-29
30-39
40-49


0.81
0.00
0.35
1.40
1.98


3.63
0.00
0.66
3.20
4.35

2.55
0.02
0.96
2.39
3.32


1.18
0.01
0.64
1.94
3.04

0.73
0.00
0.36
1.63
2.52

0.61
0.00
0.22
1.08
1.66


3.84
0.28
2.07
4.07
5.37

2.32
0.21
1.54
2.96
3.70

0.95
0.07
1.01
2.14
2.60

0.77
0.02
0.56
1.80
2.32

0.85
0.00
0.36
1.43
2.02


3.96
0.29
2.12
4.44
5.94

2.26
0.20
1.60
3.29
4.51

0.78
0.09
0.94
2.28
3.45

0.49
0.02
0.40
1.35
1.79

0.55
0.20
0.17
1.15
1.24


Source: Weighted 3% sample data of Metropolitan Sao Paulo,
Census.


1980 Brazilian









What's surprising about the distribution of income levels is that over

two thirds (68.8%) of the women ages 15-49 belong to the lowest income level,


and over five sixths (85.7%) are in the bottom two income levels.


This is a


vivid description of the labor force participation and the economic status of

the women under study here.


Table 3.5
Mean Children Ever Born to Women of 15-49 Years of Age
by Income, Age and Color Groups, Metropolitan Sao Paulo, Brazil (1980)


Income


Level


Total


Asian


White


Afro-Brazilian


1 MW
15-19
20-29
30-39
40-49


2MW
15-19
20-29
30-39
40-49


2.19
0.14
1.55
3.20
4.39


1.25
0.04
0.62
2.61
3.57


1.85
0.00
0.75
2.45
3.47

0.74
0.01
0.13
1.44
2.70


2.11
0.13
1.49
3.05
4.14

1.17
0.03
0.55
2.52
3.41


2.48
0.18
1.74
3.80
5.50

1.48
0.06
0.77
2.83
3.99


3MW
15-19
20-29
30-39
40-49


Above 3
15-19
20-29
30-39
40-49


1.19
0.04
0.44
1.97
2.95


MW


0.87
0.00
0.17
1.00
2.27


1.11
0.03
0.40
1.88
2.80


1.17
0.03
0.42
1.49


1.60
0.06
0.63
2.37
3.54


1.27
0.23
0.48
1.68
2.63









At every income level,


the fertility level for Asian is lower than that of


whites, which is in turn consistently lower than that of Afro-Brazilians.
However, the gaps between the means for Asians and those for whites in

every income group are much bigger than those between the means for

whites and those for Afro-Brazilians, indicating again that Asians are


significantly different from the other two groups,


as far as fertility is


concerned


, even when income is controlled.


More importantly,


this shows


that, after controlling for income,


three color groups.


there are still fertility variations among the


When both income and age are controlled, the mean


number of children for Afro-Brazilian women is higher than that for white


women at all income levels and in all age groups,


and the mean number of


children for white women is higher than that for Asian women at all income
levels and in all age groups.

Table 3.6 shows the fertility differences among the three color groups,


controlling for residence and age.


As expected, women in rura


areas have a


much higher fertility level than their urban counterparts.


level for rural women is 42


In fact, the fertility


more than that for urban women (2.58 for rural


women vs.


1.81 for urban women).


However, because rural women comprise


only 9.6% of the population of women, their high fertility level has little
impact on the fertility of the total population.

Color differences remain much the same in all age groups as well,


controlling for residence.


after


The pattern shown here conforms to the general


pattern exhibited by the data so far, i.e.,


average than do whites,


Asians have fewer children on


who in turn have fewer children on average than do









level for rural women, whether


Asian,


white or


Afro-Brazilian


consistently higher than that for urban women.


Table 3.6
Mean Children Ever Born to Women of 15-49 Years of Age


by Residence,


Age and Color Groups,


Metropolitan SAo Paulo,


Brazil (1980)


Residence


Total


Asian


White


Afro-Brazilian


Urban


15-19
20-29
30-39
40-49


3.83


1.07
2.60
3.60


1.34
3.28
4.84


0.42
1.85
3.08


Rural


0.17


20-29
30-39
40-49


Source: Weighted 3
Census.


Table 3.7 illu


0.18
1.66
3.81
5.34


2.94
0.17
1.95
4.59
6.45


sample data of Metropolitan Sao Paulo,


states color differentials in fertility


0.00
0.44
2.62
3.88


1980 Brazilian


controlling for both


residence and education.


Again, in both urban and rural areas, color


differences in fertility for women with no schooling are very small.


mean number of children for


Asian,


category in urban areas are 3


white and Afro-Brazilian women of this


and 3.85, respectively


the three color groups in rural areas are 4.


4.22 and 4.


r, whereas those for

, respectively.


In urban areas, the fertility level of Afro-Brazilians with any schooling
ah,"nvam nno Vosr 1C lnxAror fbsan neil- nnl'iz fhci ,I- 1'thni ,,..1r,.,if n,. ,,..n. -,,.., ,,,..4









whites at the levels of 1-4 and 5-8 years of schooling (2


and 1.21 for Asians


vs. 2.32 and 0.97 for whites).


For the top two educational level


(9-11 and 12 or


more years of schooling)


Asians; 0.


whites have slightly higher fertility level than


and 0.85 for whites and 0.73 and 0.62 for Asians.


Table 3.7
Mean Children Born to Women of 15-49 Years of Age
by Residence, Education and Color, Metropolitan Sao Paulo, Brazil (1980)


Residence


Mean


Asian (


White


Afro-Brazilian (


Urban


Zero


3.77


1.39 (100.0)
3.51 (2.3)


(21.8)


1.74 (100.0)


3.73
2.32


(8.4)
(43.5)


2.09 (100.0)
3.85 (18.7)


2.25


1 (15.5)


9-11


0.74


Rural


Zero


2.29
0.72
0.60


0.73 (21.5)
0.62 (15.6)

1.81 (100.0)
4.25 (4.9)
2.74 (45.4)
0.94 (23.1)
0.65 (19.4)


0.16 (7.3)

1.44


Total


0.97 (23.1)


(16.1)


0.85 (8.8)

2.49 (100.0)
4.22 (24.0)
2.26 (60.9)


(9.8)


0.60 (4.2)
0.87 (1.1)

1.82


0.79 (21.8)
0.48 (6.1)
0.55 (1.4)

2.94 (100.0)
4.51 (35.5)
2.35 (54.0)
0.67 (8.7)


(1.7)


0.49 (0.09)

2.18


Source: Weighted 3% sample data of Metropolitan Sao Paulo, 1980 Brazilian
Census.


Note:


The percentages in brackets are the proportions of people belonging to


various educational level


within color groups.


The patterns in rural areas are quite different; the fertility level of
Asian is higher than that of whites at all levels except at the level of 12 or









levels (about 85


of whites and 90


of Afro-Brazilians


vs. about 50


Asians),


their overall fertility levels are still higher than that of Asians.


people with no schooling at all,


the mean number of children for Afro-


Brazilians (4.51) is the highest among the three groups (4.


4.22 for whites).


for Asians and


Of those with 1-4 years of schooling, Asians have the highest


fertility level,


2.74, compared to


2.35 for Afro-Brazilians and


26 for whites.


At the levels of 5-8 and 9-11


years of schooling, Afro-Brazilians have the


lowest mean (0.67 and 0.52),


but they account for only


less than 10% of their


rural population.


The low fertility level of whites and Afro-Brazilians with


twelve or more years of schooling (0.87 for the former and 0.49 for the latter)
does not contribute much to their overall fertility level because they account


for only about 1


of their respective populations.


fertility level of Asians with 12


On the other hand


or more years of schooling (0.16),


which is


substantially lower than that for the two other groups, affects their overall


fertility level since they account for more than


of Asian in rural areas.


Considering the overall mean fertility level for each group, it appears


that there are two causes for the unpredicted distribution:


1) proportionally


Asian women are over-represented in the top two educational levels (about


46%),


compared to whites and Afro-Brazilians (about 13% and


respectively);
schooling (18.


2) Afro-Brazilians are over-represented in the category of no


in urban areas and 35


in rural areas).


Thus


the effect of


education seems to be different for the three groups.


In particular, education


seems to have greater negative impact on the fertility level of Afro-Brazilians


than on that of whites and Asians. If ti


his is true, the results here then









counterparts.


It also suggests that one's educational level is an important


factor in determining one's


fertility level,


regardless of residence and color.


Table 3.8 describes the color differentials in fertility,


controlling for


residence and income simultaneously.


As shown above


, fertility levels for all


groups in urban areas are lower than those in rural areas, and Asians have

lower fertility levels than whites in every income level, who in turn have

lower fertility levels than Afro-Brazilians, in both urban and rural areas.


Table 3.8
Mean Children Ever Born to Women of 15-49 Years of Age
by Residence, Income and Color, Metropolitan SAo Paulo, Brazil (1980)


Residence


Mean


Asian


White


Afro-Brazilian


Urban


MW


2 MW


To 3 MW
Above 3 MW


1.74
2.04


2.09
2.40


0.76
0.62
0.88


Rural


MW


2 MW


To 3 MW
Above 3 MW


2.67


2.57


0.57
0.22


2.94
3.04
1.91
3.64
2.92


Total


Source: Weighted 3% sample data of Metropolitan Sio Paulo, 1980 Brazilian
Census.


To find out the degree of association between fertility level and the


independent variables,


while controlling for some or all the other variables,


.I cn ',et- nC( tyn. tr 1n r r-n-l r erF en, ,^ ..^ 4-n' fl.4 ,- .11 1l..r k 1


'~ I n1- /Inrn n^ 4'/ ^> nlr~ rtr r tl-^I/^-









whites is considered as the reference group, against which the other two color


groups are compared.


Age and years of schooling are treated as interval


variables without modifications, but income is treated as an interval variable


with modifications such that the minimum wage in


1980 (4,150 cruzeiros) is


used as the unit of income, instead of the original unit (one cruzeiros) in the

census.

In order to compare the effects of various variables on fertility level, a


total of seven regression models are developed.


The first model measures the


effects of age and residence, the second one measures the effects of not only

age and residence but also the socioeconomic variables, education and


income.


The third model measures the effects of age, residence and color, and


the fourth one, the complete model, measures the effects of all the variables


examined here.


Models 5-7 are developed solely to examine whether


education and income have different effects on different color groups.

Based on the findings in the previous descriptive analysis, I first build a


regression model with only age and residence.


This model tells us three


things: 1) One unit of increase in age increases the mean number of children


by 0.1461,


with residence included in the model


2) being in rural areas


increases the mean number of children by 0.8294,


model with the two variables explains 36.


the tota


with age considered; 3) this


(the R-square for the model) of


variation in fertility for all the people in the sample data (see Model


(1) in Table 3.9).

To examine the cumulative effects of age, residence and socioeconomic
variables on fertility, education and income are entered into the existing









variables in Model 1; 2) when education and income are introduced into the


model


, the coefficient of age decreases slightly, but the coefficient of rural


areas (as opposed to urban areas) decreases dramatically by more than 50%,
suggesting a relatively high degree of covariation between residence,

education and income; 3) the negative signs of the coefficients of education

and income indicate a negative correlation between education and fertility,


and between income and fertility.


More specifically,


one year of increase in


schooling reduces the mean number of children by 0.2838,


and an increase of


one minimum wage in average income reduces the mean number of

children by 0.0997.


In order to measure the effects of color


, and to compare them to those


of education and income, Model 3 is obtained by adding the dummy variables
representing Afro-Brazilians and Asians (whites is the reference group) into


the first model.


There are several things to point out here:


First, unlike in


Model


, there are little changes in the coefficients for age and rural areas in


Model 3, compared to Model 1


indicating that variations in age and residence


do not contribute much to the color differences in fertility.


Second


when


Afro-Brazilians and Asians are compared to whites, they both differ
significantly from whites; the positive sign of the coefficient for Afro-
Brazilians indicates a higher fertility level than that of whites, and the

negative sign for the coefficient of Asians indicates a lower rate relative to


whites.

average,


Specifically


controlling for age and residence, Afro-Brazilians, on


have 0.4974 more children than do whites


, and Asians, on average,


have 0.6241 fewer children than do whites.


Third, a mere increase of 0.97% in









better than the first model without the color variables in explaining the total

variations in fertility for the sample data.


Table 3.9
Children Ever Born to Women Aged 20-49


Regressed on Age, Residence, Education,


Independent
Variables


Income and Color


Models


AB*


Age


.1461


.1306


.1473


.1230


.1669


.1125


Residence
Urban (Reference)


Rural


.8294


.3635


.8261


.3793


.3449


.4694


.2070


Education


-.2838


Income**


-.2720


-.0997


-.0992


-.2795

-.0851


-.2272

-.2771


-.2080

-.0809


Color
Whites (Reference)


Afro-Brazilians


Asians


.4974
-.6241


.1827
-.2476


.3674


.4351


.3771


.4364


.4382


.4400


.4996


Constant


.4024


-.6340


-2.5355


-.7518


-.4902


-1.445


-.7650


Note:


= Whites


= Afro-Brazilians


and A


= Asians


**The unit of income is the minimum wage in


P-value for all coefficients


1980 (4,150 cruzeiros).


.000.


As expected,


the fourth model


, the complete model with all the









slightly

from 43


. Meanwhile, the R-square in Model 4 increase only 0.13


p.51


in Model


0.13% more of the tota


variables in Model


to 43.64


2. This indicates that the color variables explain only

l variation in fertility that is not explained by the other

However, compared to Model 3, the coefficients of


Afro-Brazilians and Asians (as opposed to whites) drop significantly from


0.4974 to 0.1827 for the former and from -0.6241 to -0.2476 for the latter.


This


suggests that when the effects of age, residence, education, income and color

are measured simultaneously, Afro-Brazilians and Asians differ less from


whites in fertility level.


To put it differently,


most of the fertility differences


between whites and Afro-Brazilians


and between whites and Asians are due


to factors other than color.


Models 5


, 6 and 7 are developed to test the hypothesis that education


and income have different effects on different color groups.


regression analysis for each of the three group,


I run a separate


with the variables of age,


residence, education and income; Model 5 is for whites,


Model 6 is Afro-


Brazilians and Model


7 is for Asians.


Thus, we can measure the effect of the


same variable on different color group by comparing the coefficients of this
variable across Models 5, 6 and 7.


First


the coefficients of education in the three models show that


education does have different effects on the fertility levels of the three color


groups, but not in the order I expected.


Specifically


the coefficients of


education in Models


(-0.2795 for whites,


2272


for Afro-Brazilians and


-0.2080 for Asians) tell us that the (negative) effect of education is greater for


whites than it is for Afro-Brazilians. and it is the least for Asians.


In other









Second, the (negative) effect of income on fertility level is much greater
for Afro-Brazilians than it is for whites and Asians, as indicated by the


coefficients of income the three models.


We can interpret the coefficients of


income in the following way; a one unit of increase in income, i.e., an
increase of 4,150 cruzeiros, results in a reduction of 0.2771 in the mean


number of children for Afro-Brazilians


while it results in a reduction of


0.0851 and 0.0809 in the mean number of children for whites and Asians,


respectively.


Finally, the R-squares in Models 5 and 6 (0.4382 and 0.4400) are


very similar, but they are somewhat different from the R-square in Model


(0.4996).


This


indicates that the variables in the models explain


approximately the same amount of variance in fertility for whites and Afro-
Brazilians, but they explain slightly more of the variation in fertility for

Asians.


In addition


, we see the coefficient of age for Afro-Brazilians in Model 6


(0.1669) is considerably higher than those for whites (0.1125) and Asians


(0.1230).


This indicates that age has greater positive impact on the fertility


level of Afro-Brazilians than that of whites or


Asians


which confirms the


conclusion from the descriptive analysis that Afro-Brazilians have children at


younger ages than do the other two groups.


We also notice that the


coefficient of the dummy variable, rural areas, for Afro-Brazilians in Model 6


(0.4694) is much higher than that for whites (0.3449) in Model


Asians (0.2070) in Model


5 and that for


. This suggests that the gap between the fertility


level of urban and rural residents is bigger for Afro-Brazilians than it is either
for whites or Asians, which is consistent with the result of the descriptive









Summary


The descriptive analyses of the sample data show that the fertility level

of Brazilian women varies by age, color, education, income and place of


residence.


When age is controlled,


Asians have the lowest mean number of


children and Afro-Brazilians have the highest mean,


with whites in between


min every age group.
When education and age are controlled simultaneously, the existing

patterns of fertility differences among the three color groups change

completely; the fertility level of Afro-Brazilians is the lowest among the three


group, except at the level of no


are minimal.


schooling, even where the color differences


This indicates that fertility is associated with more with


education than with color.


On the other hand


, Asians have children at older


ages than do whites, and whites have children at older ages than do Afro-
Brazilians at the educational levels of less than 9 years of schooling. At


higher educational level


, it is just the opposite; Afro-Brazilians have fewer


children than do whites in all age groups, and do Asians in most age groups.

Color differences in fertility narrow a great deal when income and age


are controlled simultaneously.


The change in fertility level is most


pronounced between the first and second income level for all three color


groups.


However, in spite of the decreasing gaps among the three groups at


higher income levels, Asian fertility level is the lowest,


in the middle


white fertility level is


, and Afro-Brazilian fertility level is the highest at all age levels.


When residence and age are controlled simultaneously


c


color differences in


l









When both residence and education are controlled


there are some


interesting changes in the fertility differences among the three color groups.


In urban areas


, Asian fertility level exceeds that of whites at the educational


levels of 1-4 and 5-8 years of schooling, while the opposite is true at higher


levels for these two groups.


As before, the fertility level of Afro-Brazilians is


the lowest among the three groups, except at the level of no schooling. In
rural areas, Asians have the highest mean number of children among the

three groups, except at the level of no schooling, where Afro-Brazilians have

the highest mean. Nonetheless, the overall mean number of children for
Asians is still the lowest in both urban and rural areas due to their much
higher concentration at higher educational levels than the other two groups.

Color differences in fertility do not change much when both residence and

income are controlled.
The multivariate regression models show quantitatively the effects of


various independent variables on fertility level.


Model 1 in


Table 3.9 tells us


that both age and rural areas (as opposed to urban areas) are positively


associated with fertility level,


though the impact of the latter is much greater


and age and residence account for 36.7% of the total variation in fertility for


the sample.


The variables representing socioeconomic status, education and


income, explain 6.8% more of the variation in fertility,


with education


having greater negative impact than income on fertility (see Model 2 in


Table


3.9).


Specifically,


the mean number of children reduces by 0.2838 with one


unit (year) increase in schooling, and reduces by 0.0997 with one unit (4,150
cruzeiros) increase min average income.









different from whites in terms of fertility; on average, the mean number of
children for Afro-Brazilians is 0.4974 higher than that of whites and the mean


number of children for Asians is 0.6241 lower than that of whites


controlling for age and residence.


after


A comparison of the R-square values in


Models


2 and 3 indicates that socioeconomic status


education and


income


, has far greater impact than does color on fertility.


Model 4


, the complete model with all the variables, is very similar to


Model


to Model


, which does not include the dummy variables for color.


, the R-square increases only 0.13


Compared


in Model 4, suggesting that the


negligible effect of color on fertility


the model.


after controlling for the other variables in


However, the large decreases in the coefficients of the dummy


variables for color from Model 3 to Model 4 indicate that controlling for
education and income, the three color groups do not differ as much as they

did before these variables were controlled.

These findings support the social characteristics approach because in
general groups with higher educational attainment and higher income have


lower fertility levels.


For example, Asians have the highest educational


attainment (a mean of 6.65 years of schooling) and highest mean income
(7,261 cruzeiros) among the three group, and their fertility level is the lowest.


Likewise


, Afro-Brazilians have the lowest educational attainment (a mean of


3.23 years of schooling) and lowest mean income (3,030 cruzeiros) among the


three color groups, hence the highest fertility level.


The overall educational


level and mean income of whites rank second (4.9 years of schooling and


4,783 cruzeiros)


and therefore


, their overall fertility level is above that of









In addition, the fact that Asians differ more from whites than do Afro-
Brazilians suggests that cultural factors, such as religion and values and


norms on fertility behavior might be at work.


Unfortunately,


since the


census data do not allow us to examine the effect of cultural factors on


fertility


I can not address this issue empirically


In order to adequately


examine the complex causes of differential fertility outcomes among various
social groups, we need to conduct qualitative, as well as quantitative, research,

and consider a range of factors that are relevant to the problem.














CHAPTER 4
CHILD MORTALITY DIFFERENTIALS


AMONG


ASIANS,


WHITES AND


AFRO-BRAZILIANS


Major Determinants of Mortality and Racial/Ethnic Differentials in Mortality


It is well known in the demographic literature that the mortality rate of


a population is determined not only by biological factors (e.g.,


age,


sex and


some genetic differences) and environmental factors (e.g.,


climate and natural


resources) but also by


socioeconomic factors (e.g.,


income, education and


occupation) and cultural factors (e.g.,


group,


membership in different racial/ethnic


religious affiliation, and customs and practices related to health status).


In other words, mortality rate of a population is the result of the interplay
between the biological, environmental, socioeconomic and cultural

conditions of the society in which the population in question live at a


particular time period.


On the relationship among the above factors,


Vallin


(1980:27) pointed out that:


There is growing evidence that, within a framework of biological


constraints (progressive aging of the body,


limited life-span), and


taking into account the geographical context that may modify
these constraints, the main differences in mortality are of
socioeconomic and cultural origin.










(1) public health services, which influence mortality regardless
of individual behavior (such as spraying insecticides that control


malaria)


(2) health and environmental services that reduce the


costs of health care but require some individual responses (e.g.,


the availability of clean water)


(3) and an array of individual


characteristics (such as income, which affects health through
nutrition and housing, and education) associated with the speed
and efficiency with which individuals respond to health services
and environmental threats (1992:709).


Because mortality is the result of the interaction among these complex
factors, it is an important indicator of quality of life of a population or a sub-


population that has its distinctive characteristics.


Similarly, infant and child


mortality rates provide a summary measure of the quality of life of a

population, especially in developing countries, since they are also very


sensitive to the conditions of the above factors


endogenouss" and "exogenous"
childhood, respectively. Infant


in addition to the


causes that are particular to infancy and early
and child mortality rate is therefore used as a


fairly reliable index of social and public health conditions throughout the
world.

In modern societies that are marked by socioeconomic differences, we
see a great deal of variation among various social groups in terms of


mortality rate.


When socioeconomic differences are largely based on


racial/ethnic group affiliation, as they are in many societies, mortality


differentials vary along racial/ethnic lines as well.


For example, in the


United State, blacks have had a higher mortality rate than whites since 1940,
_. - -- n -l - 1 U 1. A i t - r _- *









analyzing race differences in adult mortality


with controls for


sociodemographic factors,


Rogers (1992) found that:


1) The demographic


variables, race,


age and


appear to be related significantly to mortality


when no other variables are controlled


, 2) when family size and marital


status or socioeconomic status is controlled separately


mortality reduces considerably


age,


racial differences in


3) when all of the sociodemographic variables,


marital status, family size and income, are controlled


simultaneously, race differences in mortality are eliminated.


Thus


it is the


sociodemographic factors,


not race itself, that are the real causes of mortality


differentials between whites and blacks in the United States.

Wood and Lovell (1992) examined racial inequality in child mortality


and life expectancy in Brazil, using the 1950 and 1980 Brazilian Census.


They


found that although the life expectancy for whites and nonwhites increased


by more than 18 and 19 years, respectively


from 1950 to 1980


them remained about the same over the 30-year period:


outlived nonwhites by


, the gap between


"In 1950


whites


years; in 1980, the comparable figure was 6


.7 years"


(1992:721)


. Farley and Allen (1989:47) reported the difference in infant


mortality between whites and blacks in the U


during 1980s:


"black children


are about twice as likely as white children to die before attaining their first


birthday."


They also described the differences in the life span between whites


and blacks in 1980; the life expectancy of white men (70


.7 years) was seven


years longer than that of blacks (63.


years),


and the life expectancy of white


women (78.1 years) was


5.8 years longer than that of black women (72.3 years).


The purpose of this chapter is to use child mortality as a measure to









children


, and (B) to find out whether differences are due exclusively to


socioeconomic standing.


If skin color continues to explain variences in child


mortality, as I expect, then the findings suggest (although do not directly test)
that cultural factors may be at work.




Child Mortality Differentials and Life Expectancy by Color Group


In this chapter, I first describe a few of the major socioeconomic
indicators for Brazilian women of the three color groups, and then measure
child mortality level of each group, using the indirect methods developed by


Brass (Brass et al.


1968) and Trussell and Preston (Trussel and Preston 1982)


(See Appendix B).


Finally,


I will analyze the association between the


socioeconomic indicators and mortality level by applying the


Tobit regression


procedure.
The sample data used here for mortality measurement of Brazilian


women consist of households with women aged 20-29,


birth.


with at least one live


The variables selected as indicators of socioeconomic status of Brazilian


women are the educational attainment of both the wife and husband,
monthly household income, participation in the social security system and

presence of piped water in the house. The importance of parental education

(especially mother's) and household income on child mortality is widely


documented in the literature


The educational attainment of the wife and


husband is here measured by the number of years of school completed by









Chapter


1 for details).


Whether or not a household participate in the social


security system is an indicator of access to public health facilities because
membership in the social security system entitles people to medical services


(Wood and Lovell 1992).


Presence or absence of running water in the house


is an important indicator of housing quality, which has a significant effect on


child mortality (Merrick 1985, Wood and Lovell 1992).


Table 4.1 shows


marked differences in socioeconomic indicators of the three groups:



TABLE 4.1
Social Indicators by Color Group, Metropolitan Sao Paulo, Brazil (1980)*

Total Afro White Asian
Social Indicator (1) (2) (3) (4)


Mother'
Father's


Education**
Education**


Household


Income***


with social Security
with Piped Water


28,773
86.8
81.1


21,588
83.3
68.5


31,276
88.3
86.4


72,227
88.9
98.0


*The data include households with women aged 20-29 years, with at least one
live birth.
**Average years of school completed
**In 1980 Cruzeiros


Afro-Brazilian
three populations. T


women have the lowest educational attainment of the


'he average years of schooling among Afro-Brazilians


(3.2) is about a year below the comparable figure for white women (4.1), and


over two years below that of Asian wnmon (3 51


Tha camo ntfmorn hnlra1c Fnr









among white households is about 45 percent higher than the income earned

by Afro-Brazilians; Asian households, on the other hand, enjoy an income
level that is 335 percent higher than that of Afro-Brazilians and 231 percent


higher than that of whites.


In contrast, the differences in the percent of


households having the social security system is the smallest among the three


color groups; 83.3


of Afro-Brazilian households, 88.3% of white households


and 88.9% of Asian households.
Because the level of mortality of a population or a subpopulation is

determined by the combined effects of such factors as income, housing,
education and access to medical care, I expect to find corresponding
differences in the survival probabilities of children born to white, Asian and


Afro-Brazilian


women.


Advances in indirect techniques of estimating the probability of death
in the early childhood years have greatly enhanced the scope and accuracy of


mortality research.


Traditional measures of the death rate rely on vital


registration statistics.


The alternative approach,


developed by William Brass


(Brass et al.


1968), measures mortality indirectly from survey or census data.


In the Brass method


, the proportion of children surviving to mothers in


different age groups (20-24; 25-29 and 30-34),


correction factor


multiplied by the appropriate


, yields estimates of the probability of death by exact ages 2,


and 5.


In the following,


I estimate child mortality level for


Asians, whites


and Afro-Brazilians, using the Brass method.


A detailed description of the


Brass method by Wood and Lovell (1992) is included in Appendix 4.1.
Table 4.2 shows the estimates of mortality among children born to









two, mortality among Afro-Brazilian children --


116 per thousand -- is the


highest of all three groups (82 and 51 per thousand for whites and Asians


respectively).

Brazilians is


In fact, the probability of death by age two among Afro-

.41 times higher than the comparable figure for white children,


2.27


times higher than the estimate for Asian children.


The same pattern


holds for the probability of death between birth and ages 3 and 5 among the


three groups, i.e.,


of whites


the mortality estimate of Afro-Brazilians is higher than that


, which is in turn higher than that of Asians.


Table 4.2 also presents eo values, the average number of years expected


at birth.


They are calculated from the three estimates of child mortality


using model life tables (e.g.,


Coale and Demeny 1983).


The eO estimates


indicate an expectation of life of 59.14 years for Afro-Brazilian,


years for


whites, and


12 years for Asians.


In other words


based on the child


mortality levels of the sample data,


Asians are expected to live 6.35 more


years than whites, who are, in turn, expected to live 6.63 more years than


Afro-Brazilians.


These measures are interpreted as the life expectancy at birth


associated with the levels of infant and child mortality estimated among the


children born to women


0 to 34 years of age who declare themselves to be of


a given skin color in the census interview.


The mortality differentials shown in


Table 4


.2 raise an important


question: If we control for the major determinants of racial inequality (the


variables presented in


Table 4.1),


do the children of Asian women continue to


experience lower death rates compared to the children born to white or Afro-
Brazilian mothers? If the mother's skin color is no longer statistically









with the mortality of her children, the results indicate that additional factors

are at work.


TABLE 4.2
Measures of Child Mortality by Color Group,
Metropolitan Sao Paulo, Brazil (1980)


Mortality Afro-
Measure* Total Braizilian White Asian


290 .116 .082 .051
390 .126 .087 .054
5%0 .134 .093 .056

e 59.14 65.77 72.12

Mortality Ratio 1.09 1.39 .96 .49


*xo0 is the probability of death between age 0 and exact age
number of years of life expected, associated with the xq0 v


life table).


The mortality ratio is the mean value of the ratio of actual to


x. e0 is the average
alues (south model


expected proportion dead among children of women with at least one live
birth.


To simultaneously control for the several independent variables it is


necessary to apply multivariate techniques.


Rather than relying on mortality


rates for groups of women by age, as in the Brass method noted above, we
need, as a dependent variable, a measure of the mortality experience for each


woman in the sample.


calculating just such a measure.


Trussell and Preston (1982) proposed a method for


Trussell-Preston technique provides a


vl~ /^








population as a whole should be 1.00.


Indeed, the estimate of the average


mortality ratio for metropolitan SAo Paulo is 1.09, as shown at the bottom of


Table 4.


Among Afro-Brazilians,


the mortality ratio is


, indicating that


the actual number of deaths substantially exceeds the expected number.


the other hand


, the mortality ratio of .96 for whites is slightly below the


expected number, and the mortality ratio of .49 for Asians is about half that of


the total population.


In effect, the mortality ratio confirms the racial


differentials in child mortality estimated in terms of


x90 and eO values in


Table 4.2.

The mortality ratio for individual women is of additional value


because it permits the use of regression analysis.


Appendix 4.1,


For the reasons discussed in


Tobit regression procedure is the appropriate in this case.


The results of regressing the mortality ratio on the various social indicators


are given in


Table 4.3.


The model refers to the population of all women 20 to


29 years of age in metropolitan Sao Paulo.


The negative signs for the


coefficients indicate that maternal and paternal levels of education reduce


mortality, as does income, membership in the social security system,


presence of piped water in the home.

The numbers given in parenthe

coefficients are measures of elasticity.


and the


Table 4.3 below the regression


On the basis of these estimates, we can


conclude that a one percent increase in mother's education reduces mortality


13.5 percent.


Father's educational attainment and household income also


reduce mortality


, respectively.


but to a lesser degree, as indicated by elasticities of


Similarly


.095 and


net of the effects of the other variables in the


,