Segmented Assimilation, Transnationalism, and Educational Attainment of Brazilian Migrant Children in Japan

Transcription

1 doi: /imig Segmented Assimilation, Transnationalism, and Educational Attainment of Brazilian Migrant Children in Japan Hirohisa Takenoshita*, Yoshimi Chitose**, Shigehiro Ikegami*** and Eunice Akemi Ishikawa*** INTRODUCTION 1 The process of international migration has recently become increasingly complex and diverse in many countries. In attempts to understand the complicated conditions for international migration, two important branches of theory have emerged: the segmented assimilation theory and theories of transnationalism. The segmented assimilation theory has emphasized that the trajectories of assimilation depend on the characteristics of the immigrants and the social environments for receiving these immigrants (Portes et al., 2005). The impact of transnationalism on educational attainment among immigrant children must also be considered, because recent technological developments in communication and transportation have enabled migrants to move regularly across borders and maintain their ties with people in the sending community (Tsuda, 2003). However, transnational activities among immigrants have not been conducted outside the local settings of immigrant communities (Itzigsohn and Saucedo, 2002). To understand the conditions for immigrants and their offspring, two important immigration theories must therefore be considered. These theoretical arguments, although useful in understanding immigrants in different settings, have been premised mainly on the experience of immigrants in the US. For instance, the model of downward assimilation in the segmented assimilation theory rests on the unique history of race relations in the US (Boyd, 2002). Widespread poverty and a high incidence of crime have been observed in American inner-city ghettos. While some immigrants in Europe reside in more deprived areas, these places have not been as seriously deprived, in terms of their social problems, as their US counterparts (Thomson and Crul, 2007). Furthermore, the segmented assimilation theory appears to neglect the national context within which immigrant children are accepted (Portes and Rumbaut, 2001). In contrast, research on immigrant children in European countries has revealed that national variations in educational systems can explain differing patterns of integration among immigrant children (Thomson and Crul, 2007). However, little is known about the integration of immigrant children in Asian countries, several of which have recently accepted large numbers of immigrants from other countries (Tsuda, 2006). The research on the educational attainment of immigrant offspring in Japan may contribute to the further development of immigration studies because the national context for the reception of immigrants differs widely between Asian and European countries (Skrentny et al., 2007). * Shizuoka University, Japan. ** National Institute of Population and Social Security Research, Japan. *** Shizuoka University of Art and Culture, Japan The Authors International Migration 2013 IOM International Migration 2013 Published by John Wiley & Sons Ltd. ISSN

2 2 Takenoshita, Chitose, Ikegami and Ishikawa In addition, the segmented assimilation theory has neglected the impact of transnational practices on the second generation of immigrants. Recent studies on immigrants transnationalism have begun to examine the impact of parents transnational activities on their offspring. Levitt (2002) highlighted the fact that socioeconomic characteristics make a difference in transnational practices among immigrant children. When the second generation of immigrants has the necessary education, language, and job skills to function effectively in multiple settings, they are more likely to choose a transnational lifestyle. Conversely, some immigrant children, through transnational practices, tend to fail to learn both the language in a country of origin and that in the host society. Frequent movement between two countries has prevented some immigrant children from acquiring sufficient language, education, and job skills, resulting in their social and economic marginalization in both societies (Levitt, 2001, 2002). However, it is unclear what effect parental transnationalism has on children s schooling in institutional settings other than North America and Europe. This study will explain the educational attainment of Brazilian immigrant children in Japan in terms of segmented assimilation and transnationalism. By observing the children of immigrants living in Japan, we can explore the impact of segmented assimilation and parental transnationalism on immigrant children in different institutional settings. As argued below, the context for the reception of Brazilian immigrants in Japan hinders, rather than promotes, upward mobility among their offspring. In other words, downward assimilation seems to be prevalent among Brazilian migrant children in Japan. CONTEXT FOR THE RECEPTION OF BRAZILIAN IMMIGRANTS AND THEIR OFFSPRING IN JAPAN The immigration policy and labour market structure in the destination country form very important institutions, strongly affecting the process of assimilation among immigrant children (Portes et al., 2005). The change in ethnic return migration policy in Japan should be noted here, as it has been a major factor in the rapid increase in the number of Brazilian migrants in Japan since The revised immigration policy enacted in 1990 permitted second- and third-generation Nikkeijin, descendants of Japanese emigrants, to enter Japan without any restrictions on their activities (Tsuda, 2003). According to the foreign registration statistics, although only 14,000 Brazilians lived in Japan in 1989, the population of Brazilians soon exceeded 100,000 in 1991 and was over 200,000 in 1996, finally reaching approximately 316,000 in This revision of the immigration law also enabled Brazilian migrants to move back and forth between Brazil and Japan quite easily. Many Brazilian migrants consider themselves target earners rather than permanent settlers in Japan. Although some have returned to Brazil after saving a considerable sum of money earned in Japan, they also tend to face difficulties in readapting economically to Brazil after returning from Japan, due to a lack of employment opportunities in Brazil and the higher likelihood of small business failure (Tsuda, 2003; Yamamoto, 2010). Even if they return to Japan again, they must often engage in temporary employment controlled by labour brokers. These transnational practices can cause Brazilian adult immigrants to become socioeconomically marginalized in both settings (Takenoshita, 2010). How does this frequent movement between these two countries influence their children s lives? Prior ethnographic studies have suggested that the back-and-forth movement between Brazil and Japan has a negative influence on the educational attainment of Brazilian children in Japan because it hinders them from learning both Japanese and Portuguese and from acquiring the education and job skills necessary to succeed in either setting (Ishikawa, 2005). In addition, many Brazilian immigrants continue to identify themselves as temporary migrants, even after long-term residence in Japan. Under these circumstances, Brazilian children often feel puzzled over whether

3 Segmented assimilation, transnationalism, and educational attainment 3 they should study hard in Japan or prepare to return to Brazil (Ishikawa, 2005; Takenoshita, 2005). An evaluation of the labour market structure in the receiving country is also important for understanding the incorporation of immigrants. Unlike the situation in the US, bifurcation of the labour market in Japan exists between standard and nonstandard workers. Lifetime employment, seniority earnings, and firm-based labour unions have shaped the employment protections for regular workers in Japan (Koike, 1988). However, increasing global economic competition requires labour market flexibility, leading to rapid growth in the number of nonstandard workers (Genda, 2005). 3 In fact, many Brazilian workers are incorporated into the temporary employment sector controlled by labour brokerage agencies (Higuchi and Tanno, 2003). If companies use labour dispatchers, they can easily reduce the flow of labour in an economic downturn (Genda, 2005). This implies that temporary Brazilian workers must accept employment instability. Furthermore, they have unskilled jobs in the manufacturing industry that do not require any work-related skills, knowledge, or even Japanese language proficiency. The educational system should also be considered in the analysis of the national context for the reception of immigrant children (Thomson and Crul, 2007). In the Japanese educational system, compulsory schooling includes six years of elementary school and three years of middle school. Whereas education through the junior high school level is comprehensive and non-selective, senior high schools are selective and typically classified into two different tracks: academic and vocational (Ishida, 2007). Because high school is not compulsory, the Japanese government is not required to provide public education for all who desire it. Public high school admission policies are strictly based on entrance examination scores and on grades (Amano, 1996), requiring linguistic adaptation or acculturation among immigrant children. In addition, Brazilian immigrant children face great difficulty in learning the Japanese language. The written Japanese language is composed of Hiragana, Katakana, and Chinese characters (Kanji). Generally, elementary school students are required to memorize many Chinese characters, nearly 1000, in Japanese language classes over six years of schooling (Miyajima and Ohta, 2005). The Japanese linguistic system makes it more difficult for immigrant children to succeed in school (Sakuma, 2006). Let us describe the discrimination against immigrants in Japan. Tanabe (2009) suggested that the Japanese tend to prefer people from Western countries, such as the US and European countries, to people from Asia and other regions. Although Japan has been classified as an Asian country, it has sought to identify itself with Western countries, affecting Japanese perceptions of immigrants in Japan. Accordingly, immigrants who appear similar to Westerners may be less discriminated against by the Japanese than others (Yoshino, 1992). Conversely, other scholars have argued that the myth of ethnic homogeneity has also persisted in the postwar period, creating discrimination against foreign migrants regardless of their racial background (Lie, 2001). In this case, ethnicity, nationality and culture are more important than race for understanding discrimination against immigrants in Japan. For example, discrimination against the Korean minority in Japan has been based upon ethnicity rather than race (Fukuoka, 2000). Brazilian children have also suffered from prejudice and discrimination by Japanese children and teachers in public schools in Japan. Furthermore, some Brazilian children, if possible, have tried to hide their Brazilian backgrounds at school and behave as if they were Japanese to avoid being bullied by their Japanese peers (Takenoshita, 2005; Tsuda, 2003). Unfavourable reception in Japanese public schools seems to force some children of immigrants to drop out of even compulsory schooling or hinder them from enrolling in high school (Sakuma, 2006). Furthermore, it is worth mentioning the local contexts for the reception of immigrant children in Japan. The local context is important due to the lack of integration programs for immigrant children at the national level in Japan. In such a context, it has been left to the local governments to guarantee educational opportunities for immigrant children (Tsuda, 2006), leading to substantial differences in the educational policies of compulsory schooling targeted at immigrant children across

4 4 Takenoshita, Chitose, Ikegami and Ishikawa municipalities (Chitose, 2008; Sakuma, 2006). In the Shizuoka prefecture, the number of Brazilian immigrants also varies widely across municipalities. Hamamatsu, a large industrial city with a high concentration of automobile manufacturers, has attracted a large Brazilian population in Japan. The Hamamatsu municipal government has provided a variety of integration programs for immigrants and their children since the 1990s. 4 For example, in 2001, the Hamamatsu municipal government has hired many Portuguese-, Spanish- and Chinese-speaking tutors to help immigrant children who are not fluent in Japanese and provide some support to their parents. Furthermore, to facilitate the education of immigrant children, the Hamamatsu government has subsidized the funding of private ethnic schools and nonprofit organizations (NPOs) that help educate immigrant children. NPOs and ethnic schools have also played key roles in providing education for immigrant children (Tsuda, 2006). For example, one NPO has conducted several activities that assist immigrant children in learning school subjects that are important for high school entrance examinations (Takenoshita, 2005). Private ethnic schools for Brazilian immigrant children have also been established to help Brazilian children in Japan prepare for their return to Brazil 5 (Ishikawa, 2009). However, educational degrees in these ethnic schools may not allow them to obtain good jobs in the Japanese labour market or to attain higher education in Japan because the curricula at these schools do not comply with the one specified by the Japanese Ministry of Education (Sakuma, 2006; Ishikawa, 2009). ANALYTICAL STRATEGIES This study examines the influences on high school enrolment, among children of Brazilian immigrants in Japan, of factors drawn from the theories of both segmented assimilation and transnationalism. We focus on high school enrolment as a dependent outcome among Brazilian children because a large number of Brazilian children cannot enter high school, even though almost all Japanese children can do so. Moving to factors affecting the schooling of immigrant children, family background plays an important role in enabling educational progress. Parental educational attainment and socioeconomic status have been important predictors for schooling outcomes among immigrant children (Portes et al., 2005). In the context of labour market incorporation among Brazilian migrants in Japan, the father s employment status is extremely important in creating a financially secure family life. We hypothesise that Brazilian children whose fathers enjoy employment stability, although there are very few, are more likely to be enrolled in high school. Conversely, it is unclear whether having an employed mother helps or hinders immigrant children s education (Ishikawa, 2005; Chitose, 2008). In this study, we also examine the effects of the mothers work on their children s schooling. In addition, parental linguistic assimilation is worthy of attention. Portes and Rumbaut (2001) suggested three possible acculturation patterns: consonant, dissonant, and selective acculturation, thereby affecting intergenerational relationships in an immigrant family. To understand these acculturation patterns, we need to measure children s linguistic abilities. However, due to a lack of available data, this study focuses on the consequence of consonant acculturation among Brazilian migrant families by investigating the effect of Japanese language fluency among Brazilian migrant parents on their children s schooling. In general, immigrant children tend to become acculturated more quickly than their parents. Consonant acculturation occurs when parents possess sufficient human capital. Role reversal is less likely to occur when Brazilian parents speak and understand the Japanese language well because their language proficiency enables them to support their children. 6 As for other important factors, the race, gender, and age at migration of immigrant children also captured our attention. Brazilian migrants in Japan seem to be highly diverse in racial composition

5 Segmented assimilation, transnationalism, and educational attainment 5 due to a high rate of intermarriage in Brazil (Tsuda, 2003). Some Brazilian children can hide their Brazilian backgrounds to avoid being bullied by their Japanese peers because they have Nikkei parents and they look Japanese. On the other hand, children with at least one parent without Japanese ancestry may appear racially different from other Japanese. Accordingly, these children may be discriminated against by other Japanese, leading to a lower likelihood of high school enrolment. Conversely, Brazilian children who appear similar to most Japanese may still suffer from discrimination because other ethnic minority groups, such as Koreans, have experienced discrimination from the Japanese despite their similar racial backgrounds. To identify children s races in this study, we use information on the ancestry of their parents. Children with at least one parent without Japanese ancestry are considered phenotypically different from other Japanese. 7 The likelihood of high school enrolment is compared between these children and others born to parents of Japanese ancestry. Recent studies in the US have focused not only on racial and ethnic differences in educational and occupational attainment but also on gender discrepancies. In recent decades, women in the US have surpassed men in educational attainment. This change has also led to female advantages in education among some immigrant children, even though their parents come from societies with traditional gender roles (Feliciano and Rumbaut, 2005). In contrast, female disadvantages in educational and occupational attainment have persisted in Japan (Ishida, 2007). Additionally, gender gaps in schooling have been reported in Brazil, although male advantages in education have been reduced recently (Torche, 2010). Given that the female advantage in schooling among immigrant children reflects a larger phenomenon of gender equalization in US education, a female advantage in educational attainment is less likely to occur among Brazilian children in Japan due to the educational inequality present in Japan. In other words, Brazilian girls in Japan may be less likely to be enrolled in high school than Brazilian boys. Age at migration is also important because the socialization process in families, schools, and neighbourhoods depends largely on the timing of migration (Cahan et al., 2001). Children s educational pathways depend on whether or not they have learning experience in the origin country (Zhou, 1997). However, it is unclear what effect the timing of migration has on children s schooling among Brazilian children in Japan. In this study, we assumed that the younger a child is at the time of migration to Japan, the more time the child can spend adapting to the school environment and educational system in Japan (Ishikawa, 2009). Local contexts for immigrant reception may play an important role in making a difference in high school enrolment among Brazilian immigrant children in Japan, largely because the responsibility for providing education to immigrant children has been left to local governments. Private ethnic schools for Brazilian children and NPOs are also important in securing these children s educational opportunities. Nevertheless, there are regional variations in the number of organizations within the Shizuoka Prefecture, the region on which this study is mainly focused. Within the Shizuoka Prefecture, a large number of Brazilian migrants are found in Hamamatsu. This municipal government began to cope with the problems faced by Brazilian immigrants living there earlier than other municipal governments. There are more NPOs helping immigrant children and Brazilian ethnic schools there due to the large Brazilian population. To identify the local context for their reception, we compare the likelihood of high school enrolment between those Brazilian immigrant children who live in Hamamatsu and those who do not. The effect of transnationalism should be considered in the case of Brazilian migrants in Japan. This study uses the following two indicators: the experience of back-and-forth movement between Brazil and Japan and immigrant parents self-identification as temporary migrants. As argued above, many Brazilian immigrants engage in back-and-forth movement between Brazil and Japan, and they continue to identify themselves as temporary migrants rather than permanent settlers. Existing ethnographic studies have highlighted that these transnational practices among Brazilian immigrant parents hinder their children s educational attainment in Japan. This study examines the effect of parental transnationalism on children s schooling.

6 6 Takenoshita, Chitose, Ikegami and Ishikawa DATA AND MEASUREMENT The data used in this study were derived from three different sources: a survey of Brazilian immigrants in the Shizuoka Prefecture conducted in 2007, a survey of Latin American immigrants in Hamamatsu city conducted in 2006, and a social stratification and mobility (SSM) survey conducted in The first two surveys, conducted by the Shizuoka prefectural government and the Hamamatsu municipal government, respectively, aimed to investigate the extent to which these immigrants were integrated into the Japanese society. 9 All the authors of this paper assisted in conducting these two surveys. The surveys contain rich information on demographic characteristics, labour market positions, Japanese language fluency, and the education of immigrant children. The survey of Latin American immigrants in 2006 also included Peruvian and other Latin American immigrants in Hamamatsu city, whereas the 2007 survey did not. Only Brazilian samples were used in this study. These two surveys mainly employed a systemic sampling method. Aside from the samples selected at random from the Alien Registration records, supplementary samples were selected by a non-systemic sampling method to increase the number of samples in these two surveys. To obtain supplementary samples in the survey of Latin American immigrants in 2006, we asked public elementary schools and Brazilian ethnic schools in Hamamatsu to distribute the questionnaires to the Brazilian parents of children enrolled in these schools. Furthermore, we asked several labour dispatchers to distribute the questionnaires to their Brazilian workers in this survey. To obtain supplementary samples in the 2007 survey of Brazilian migrants, we also asked public schools in the Shizuoka Prefecture, which had accepted a substantial number of Brazilian children, to distribute the questionnaires to Brazilian parents. The respondents to these two surveys were originally Brazilian and other Latin American immigrants aged 16 years or older. Although combining the two datasets yielded 3,174 respondents, there were very few Brazilian respondents whose children had attended compulsory schooling in Japan and presumably graduated from it. To obtain a sizable number of samples, we decided to use the supplementary samples in this study. The total sample size in this study was thus 203. While almost 40 per cent of the samples were drawn randomly from a foreign registration list, others were chosen from the supplementary samples. Because it might give rise to non-sampling errors in our investigation, we need to cautiously evaluate those results estimated by multivariate analyses. To combine the two data sets, we considered how each item in the respective surveys was measured. In this case, there were almost no problems in combining these two data sets because the same research designs were used, and almost the same questions were asked in both surveys. Furthermore, we note that Hamamatsu city is located within the Shizuoka Prefecture. Combining the two data sets thus led to an overrepresentation of children of Brazilian migrants living in Hamamatsu city. The combined data were therefore weighted according to the proportion of Brazilian population in Hamamatsu compared with those in other municipalities in the Shizuoka Prefecture. Because we examine the transition into high school, the samples of Brazilian children used in this study were restricted to those aged between 16 and 30 who migrated to Japan at the age of 14 or below or who were born in Japan. 10 The unit of observation in this study was a child. Parental information derived from the two surveys was appended to each child s data. 11 To investigate the hypotheses drawn from the segmented assimilation and transnationalism theories with respect to children s schooling, we estimate high school enrolment using the logistic regression model. The dependent variable is a dichotomous indicator consisting of whether or not the child had a successful transition into high school. Note that high school here consists of Japanese public and private high schools as well as Brazilian ethnic schools. The set of independent variables, including the characteristics of the family and the children themselves, are shown in Table 1.

7 Segmented assimilation, transnationalism, and educational attainment 7 TABLE 1 DEFINITION OF VARIABLES Dependent Variables High school enrolment = 1 if children had successful transition into high school Independent Variables Gender = 1 if a child is a female Race Race among Brazilian children is judged, according to the information on Japanese ancestry among their parents. = 1 if a child has a parent without Japanese ancestry Father s education Secondary educated = 1 if father completed secondary education Tertiary educated = 1 if father completed tertiary education Mother s education Secondary educated = 1 if mother completed secondary education Tertiary educated = 1 if mother completed tertiary education Father s standard employment = 1 if father is in standard employment or self-employed Mother s employment = 1 if mother is working Parental Japanese fluency It was measured by five subjective scores such as reading, writing, and conversation. A combination of these scores is used as a scale for Japanese language proficiency of parents. Children s age at migration 5to9 = 1 if a child migrated into Japan at the age of 5 to 9 More than 10 = 1 if a child migrated into Japan at the age of 10 or above Having experience back and = 1 if parents have experienced back-and-forth movement between forth movement Brazil and Japan Prospect for future residency Return to Brazil in the future = 1 if parents intend to return to Brazil in the future Uncertain future prospect = 1 if parents do not decide whether they will return to Brazil or not Residence in Hamamatsu = 1 if parents live in Hamamatsu DESCRIPTIVE STATISTICS Table 2 lists the relevant background information on Brazilian immigrants in Japan by presenting the data from the Survey of Brazilian migrants in the Shizuoka Prefecture in 2007 and the SSM survey in Comparing the educational attainments of the Brazilian and Japanese samples, we observed that the share of Japanese respondents having a tertiary education was much higher than that of the Brazilian samples. For instance, only 15% of Brazilian male workers had a tertiary education, whereas 31% of Japanese male workers had a tertiary education. Previous studies have reported that Japanese Brazilians in Brazil have climbed the socioeconomic ladder through educational attainment (Tsuda, 2003). Nevertheless, we noted that the degree to which education has expanded in the postwar period differs remarkably between the two nations. According to the Brazilian census of 1991, in the state of Sao Paulo, while only 5% of native-born white people had a tertiary education, 17% of native-born people of Asian descent, most assumed to be second- or third-generation Japanese Brazilian immigrants living in Brazil (Dwyer and Lovell, 1990), had a tertiary education. This difference highlights the fact that, although Japanese-Brazilians living in Brazil have a higher educational attainment than the native-born white Brazilians, their educational level is much lower than that of the Japanese population. With respect to occupation, most Brazilian migrants in Japan have manual labour jobs. The Japanese census in 2005 showed that approximately 90% of Brazilian workers in Japan were engaged in manual labour jobs, compared with only 30% of Japanese workers. Table 2 also indicates that almost 85% of Brazilian workers are nonstandard workers, regardless of their gender. This figure for Japanese male workers is roughly 10%, whereas approximately half of Japanese female workers are nonstandard workers. The results for mean weekly working hours demonstrates that Brazilian

8 8 Takenoshita, Chitose, Ikegami and Ishikawa TABLE 2 DESCRIPTIVE STATISTICS OF BRAZILIAN AND JAPANESE WORKERS Male Female Brazilian Japanese Brazilian Japanese Education (percent) Compulsory or less Secondary Tertiary Percent having education in Japan Employment Status (percent) Standard employment Nonstandard employment Self-employment Mean weekly working hours N Note: The data for Brazilian workers are drawn from the survey for Brazilian migrants in Shizuoka Prefecture. The data for Japanese workers are based upon the social stratification and mobility survey in men and women work more hours than their Japanese counterparts, implying that Brazilian parents spend less time monitoring or controlling their children. Table 3 lists the descriptive statistics for the variables used in the following multivariate analyses and also compares children who were successfully enrolled in high school with those who were not. These results highlight the significant gap in high school enrolment between Brazilian and Japanese children. We found that 71 per cent of Brazilian children were successfully enrolled in high school, whereas almost all Japanese children were enrolled. This table also shows a striking difference in the distribution of female children between the enrolled and the unenrolled across nationality. In the Japanese sample, the share of female children among the enrolled was higher than that among the unenrolled. In contrast, in the Brazilian sample, the share of female children was higher among the unenrolled (57%) than the enrolled (41.3%), implying that Brazilian female students are in a disadvantaged position. The share of children with non-nikkei parents was higher among the unenrolled children than among the enrolled children. Their parents educational attainment reveals that this factor was associated with the children s enrolment status for both Brazilians and Japanese. There were also substantial differences between the enrolled and the unenrolled in the fathers employment status. Roughly 20 per cent of the enrolled children had a father involved in regular employment or self-employment, while the corresponding figure was only 3.7 per cent among the children who were not enrolled. It appears that the mother s employment is also positively correlated with their children s enrolment in high school. The average score for Japanese language fluency among parents of enrolled children was higher than the corresponding score for the parents of unenrolled children. With respect to the age of migration, the results in Table 3 clearly show that the share of unenrolled children was quite high for those who migrated at the age of 10 or older. This result indicates that a younger age of migration leads to a higher likelihood of high school enrolment. Residency in Hamamatsu, where many Brazilian immigrants are concentrated and where a variety of integration programs exist, was also associated with children s enrolment in high school. Among the enrolled children, the share of those residing in Hamamatsu was higher than that among the unenrolled children. With regard to transnationalism, the share of children whose parents experienced back-and-forth movement between Brazil and Japan or whose parents intended to return to Brazil was higher among the unenrolled than among the enrolled children.

9 Segmented assimilation, transnationalism, and educational attainment 9 TABLE 3 DESCRIPTIVE STATISTICS OF INDEPENDENT VARIABLES AMONG BRAZILIAN AND JAPANESE CHILDREN Brazilian Japanese Enrolled Not enrolled Total Enrolled Not enrolled Total the share of female (%) the share of those with non-nikkei parents (%) Father s educational attainment Compulsory educated (%) Secondary educated (%) Tertiary educated (%) Mother s educational attainment Compulsory educated (%) Secondary educated (%) Tertiary educated (%) Father s employment: regular workers (%) Mother s employment: being employed (%) Parental Japanese language fluency (mean) Children s age at migration native-born or less than 4 years old (%) 5 to 9 years old (%) more than 10 years old (%) Having experience of back-andforth movement (%) Prospect for future residency Return to Brazil in the future (%) Uncertain future prospect (%) Residence in Hamamatsu (%) N % These descriptive results have thus far revealed that family background, the characteristics of the children themselves, the place of residence, and parental transnationalism appear to be closely linked with high school enrolment among Brazilian immigrant children in Japan. Using logistic regression, we estimated the impact of variables drawn from the segmented assimilation and transnationalism theories on high school enrolment among these children, while taking into account the effects of other attributes. MULTIVARIATE ANALYSES Table 4 presents the results of the multivariate analyses, which compare the effects of gender and fathers education levels on the logarithmic odds of enrolment in high school between Japanese and Brazilian samples. Surprisingly, there was a substantial difference observed in the impact of gender between these two groups. There was no significant gender effect on enrolment in high school among the Japanese samples, whereas male Brazilian students were 2.5 times more likely to

10 10 Takenoshita, Chitose, Ikegami and Ishikawa TABLE 4 THE EFFECT OF GENDER AND FATHER S EDUCATION ON ENROLMENT INTO HIGH SCHOOL AMONG JAPANESE AND BRAZILIAN SAMPLES Brazilian Japanese Coef. s.e. Odds Coef. s.e. Odds Gender (Female=1) * Father s education (Reference: Compulsory educated) Secondary educated ** Tertiary educated ** ** Constant ** ** N X ** ** Pseudo R p <.10; *p <.05; **p <.01. enrol in high school than female students. We discuss the reason for this in more detail later. It was additionally shown that the impact of the father s education was considerably lower among the Brazilian samples than among the Japanese samples. This difference may be due to the fact that almost all the Brazilian parents acquired their education in Brazil. This implies that the educational attainment of immigrant offspring should be explained by other attributes. Table 5 presents the results of high school enrolment among children of Brazilian immigrants. The effect of gender on high school enrolment persisted even after other attributes were controlled for. In contrast, children s schooling did not depend on whether or not children were born to two Nikkei parents, which was considered in this study to represent a racial difference. Nevertheless, we note that only differences in children s schooling among Brazilian immigrants were explored using this race variable. This result also indicates that Brazilian immigrant children are disadvantaged, regardless of their race. With respect to educational attainment among immigrant parents, the father s education had a significant effect on the children s schooling, whereas the mother s education did not. In contrast, it was revealed that both the father s and mother s employment were associated with greater enrolment of their children in high school. Children whose fathers were involved in regular employment and whose mothers also worked were more likely to be enrolled in high school. In other words, the incorporation of Brazilian migrants into the unstable employment sector hinders their children s schooling. The mother s economic contribution also plays a positive role in their children s schooling. Furthermore, parental Japanese language proficiency significantly facilitates children s enrolment in high school. Turning to the effect of age at migration, we observed that it had a significant impact on the children s high school enrolment. Those who migrated to Japan at four years old or below and those who were born in Japan were more likely to enrol in high school. The segmented assimilation theory also places importance on the context for immigrant reception. The findings also suggest that Brazilian children who live in Hamamatsu, where a sizable number of Brazilian immigrants are concentrated and more integration programs are implemented, were more likely to enrol in high school than those residing in other cities. The results of our analyses highlight the fact that the characteristics of the children s families and the local context for their reception play decisive roles in enabling educational advancement in the case of Brazilian immigrant children in Japan.

11 Segmented assimilation, transnationalism, and educational attainment 11 TABLE 5 THE DETERMINANTS OF HIGH SCHOOL ENROLMENT AMONG CHILDREN OF BRAZILIAN MIGRANTS Coef. s.e. Odds Gender (Female=1) * Race Father s education (Reference: Compulsory educated) Secondary educated Tertiary educated * Mother s education (Reference: Compulsory or less) Secondary educated Tertiary educated Father s standard employment Mother s employment Parental Japanese fluency ** Children s age at migration (Reference: less than 4 years old) 5 to 9 years old More than 10 years old * Having experience back and forth movement ** Prospect for future residency (Reference: permanent residency in Japan) Return to Brazil in the future Uncertain future prospect Residence in Hamamatsu ** Constant N 203 X ** Pseudo R p <.10; *p <.05; **p <.01. With respect to the impact of transnationalism, the results indicated that back-and-forth movement between Brazil and Japan among the family members significantly affects children s high school enrolment. Brazilian children whose parents had no experience of moving back and forth were four times more likely to enrol in high school. As argued previously, this is partly because moving back and forth hinders Brazilian children in improving their Japanese proficiency, which would lead them to do poorly in school. In contrast, the parents prospects for returning to Brazil had no significant influence on their children s schooling, even though the coefficients showed that there may be a negative association between them. CONCLUSIONS The current research investigated the educational pathways among Brazilian immigrant children in Japan from the perspectives of segmented assimilation and transnationalism. Our research shows that Brazilian immigrant families play a major role in enabling their children s high school enrolment. Notably, the family s economic resources facilitate their children s enrolment in high school. In other words, Brazilian children s schooling is impeded by employment instability among their parents. In addition, enrolment is also conditional on the acculturation patterns among Brazilian parents. Immigrant parents who speak Japanese fluently can maintain authority over their children by communicating well with and providing some support to them. Age at migration was also associated with the

12 12 Takenoshita, Chitose, Ikegami and Ishikawa acculturation patterns among immigrant children. This study also revealed that the 1.5 generation, those who migrated to Japan at the age of 10 or older, tend to face more difficulties in high school enrolment. They do not have sufficient time to adapt to the school environment in Japan. In contrast, race does not affect immigrant children s schooling in Japan when other attributes are taken into account. The results showed that children with both parents of Japanese ancestry were not advantaged in high school enrolment. Given the context of discrimination against Brazilian immigrant children in Japan, they may be discriminated against by other Japanese peers, regardless of their racial background. In the US, discrimination against black people constitutes a barrier on the path to occupational mobility and social acceptance. In Japan, the myth of ethnic homogeneity has been so prevalent that many foreign migrants have not been totally accepted in the host society. In other words, Japan has denied identifying itself as a country of immigration, excluding immigrants from mainstream Japanese society. These arguments appear to represent the Japanese context of segmented assimilation, as argued below. The findings on gender and education are also interesting because they revealed that Brazilian girls were less likely to enrol in high school. This result shows a significant contrast with their US counterparts. Given that the female advantage in schooling among immigrant children reflects a larger phenomenon of gender equalization in US education, the lower educational attainment among Brazilian girls in Japan is probably influenced by gender disparity in educational and occupational attainment in Japan. Brazilian parents allocate a limited amount of resources to sons rather than daughters due to a gender gap with respect to the expected returns on investments in their education. In addition, several ethnographic studies have suggested that there may be greater expectations on daughters to provide care for other siblings (Ishikawa and, 2005; Takenoshita, 2005). 12 This result also highlights the fact that the institutional context of gender discrimination in Japan is also important for understanding the schooling of immigrant children. Turning to the impact of the local context for immigrant reception, the findings of this study indicate that this factor plays an important role in facilitating children s schooling. Nevertheless, the local context of reception differs widely across countries. The segmented assimilation theory in the US sheds light on the role of the ethnic community in children s schooling by observing its divergence across ethnic groups (Portes et al., 2005). Conversely, this study suggests that local governments, NPOs, and Brazilian ethnic schools play crucial roles in securing education for immigrant children. The role of the ethnic community is minimal in providing support for Brazilian children because employment instability among Brazilian immigrants has hindered them from forming their own ethnic community (Kajita et al., 2005; Takenoshita, 2010). Applying the segmented assimilation theory to other settings, we should consider cross-national differences in the local context for the reception of immigrants. With respect to transnationalism, the experience of back-and-forth movement between Brazil and Japan prevented immigrant children from enrolling in high school, whereas the parental intention to return to Brazil had no significant impact on children s schooling. In other words, this study found no evidence that parental transnationalism facilitates children s schooling. Previous literature on transnationalism in the US suggests that the consequences of transnationalism depend considerably on the family s socioeconomic background and the modes of incorporation (Levitt, 2002; Itzigsohn and Saucedo, 2002). Brazilian immigrants are incorporated into the lower segment of the labour market, and these incorporation mechanisms have increased their tendency for temporary migration. The reason for their migration to Japan is to save and remit as much money as possible rather than to provide a Japanese education for their children or to settle in Japan. Under these circumstances, it is quite understandable that the transnationalism of immigrant parents has hindered their children s schooling. This study also emphasizes that the impact of the transnationalism of immigrant parents on their children is highly dependent upon its context. To summarize these findings theoretically, we discuss the Japanese model of segmented assimilation. This study demonstrates that the model of downward assimilation proposed in US immigration

13 Segmented assimilation, transnationalism, and educational attainment 13 studies can well explain the conditions for schooling among Brazilian immigrant children in Japan. This seems to be linked to the fact that Japan has not officially identified itself as a country of immigration. Both the government and Japanese society regard foreign migrants as temporary residents rather than permanent settlers, excluding them from mainstream Japanese society. Social exclusion of immigrants in Japan is reflected in their labour market incorporation, by which Nikkei immigrants from Latin American countries have been employed in unskilled jobs with great employment instability and no opportunity for advancement. In contrast, the US is considered a classic country of immigration. Immigration in the US has shaped a fundamental part of its national identity. Accordingly, integrating immigrants into the mainstream is a very important formative factor in American society. However, there has apparently been no debate at the national level in Japan about how the society should accept and integrate immigrants. The prevalence of downward assimilation in Japan reflects the lack of discussion on how to integrate immigrants into society. This lack of an integration model in Japan has resulted in immigrant assimilation patterns less diverse than their US counterparts. The American model of segmented assimilation emphasizes the role of the ethnic community in upward mobility. However, the Japanese context for immigrant reception prevents Brazilian immigrants from forming an ethnic community. The prevalence of downward assimilation and the difficulty of upward mobility are rooted in the attitude of Japanese society toward immigrants, that is, their reluctance to accept immigrants as legitimate members of the larger society. Finally, it is worth noting that the following problems remain unresolved in this study. First, this study used supplementary samples, which were not randomly drawn from the population, to secure a sufficiently large number of samples. This may have led to some biased coefficients in our findings. Future studies should re-examine the validity of our results using other data. Second, the surveys of Brazilian immigrants we used originally targeted adult immigrants, rather than 1.5 or the new second generation. This limitation made it difficult to explore the complex patterns of acculturation among immigrant children. To resolve this problem, we should conduct a survey that targets immigrant children in a future study. Third, we attributed significant differences in children s schooling across municipalities to the role of municipal governments and NPOs in providing integration programs for them and the role of ethnic schools in securing children s education. This study, however did not examine what factors in the local context for immigrant reception might help or hinder immigrant children s education. Future studies should do so by including the institutional variables in the model and comparing their impacts on schooling outcomes across a large number of localities. NOTES 1. We thank two anonymous reviewers for their insightful comments. We are also grateful to the Shizuoka Prefectural Government and the Hamamatsu Municipal Government for allowing us to use the data drawn from the surveys of immigrants living in these places. 2. The Brazilian population in Japan has declined rapidly since the economic crisis in In 2009, it was 267,456, a 16 per cent decrease compared with the Brazilian population in Employers have provided standard workers with opportunities for skill enhancement, fringe benefits, and seniority earnings, while they have not provided the same for contingent workers. In fact, unskilled and menial jobs are highly concentrated in the nonstandard employment sector. 4. This information was mainly derived from interviews with workers in the Hamamatsu municipal government and several staff members at the Hamamatsu Foundation for International Communications and Exchange (HICE) and from the website of the Hamamatsu municipal government (

14 14 Takenoshita, Chitose, Ikegami and Ishikawa 5. The economic crisis that began in the US in 2007 also affected the Japanese economy, leading to a rapid increase in unemployment among Brazilian migrants in Japan. This crisis also made it more difficult for Brazilian parents to enrol their children in Brazilian ethnic schools in Japan. 6. Due to a lack of information on children s language abilities available to the authors, this study does not shed light on whether selective acculturation occurs among Brazilian children. 7. Race is generally classified into three groups: white, black and Asians. Please note that when using information on parental ancestries, we cannot distinguish between white and black children. 8. Because the SSM survey contains rich information on labour market positions, educational trajectories, and family background, these survey data are suitable for comparisons between Brazilian immigrant children and native Japanese children in Japan. One of the authors was allowed to use this data set. 9. All the authors of this paper assisted these two local governments in conducting the surveys of Brazilian immigrants. We created research designs for these surveys, analysed data and submitted survey reports to these governments. 10. To compare the Brazilian and Japanese samples, the Japanese samples were limited to those born between 1976 and 1985 the youngest 10-year birth cohort in the SSM survey to account for the differences in the birth cohort between the two samples. Conversely, the Brazilian samples used in this study were children born between 1978 and The target respondents in the Shizuoka and Hamamatsu surveys were immigrants 16 years of age and older. Information on children s schooling was derived from the answers of their parents, who were the respondents in both surveys. To explore the determinants of schooling among immigrant children, the unit of observation was set as one child. 12. In addition, gender gaps in children s schooling may have resulted from the fact that Japanese Brazilians in Brazil treat sons and daughters differently. Japan has been characterized as a country with a strong gender division of labour. Given the time of their emigration from Japan, it is understandable that many Japanese immigrants in Brazil held traditional gender stereotypes. Yamamoto and (2010) reported that unmarried young Japanese Brazilian women migrated to Japan because they wanted to become economically independent. In contrast, their Brazilian parents in Brazil urge their daughters to get married because, given a typical gender stereotype, women can rely economically on their husbands. REFERENCES Amano, I Nihon no Kyoiku Shisutemu: Kozo to Hendo, University of Tokyo Press, Tokyo. Boyd, M Educational attainments of immigrant offspring: Success or segmented assimilation?, International Migration Review, 36(4): Cahan, S., D. Davis, and R. Staub 2001 Age at immigration and scholastic achievement in school-age children: Is there a vulnerable age?, International Migration Review, 35(2): Chitose, Y Compulsory schooling of immigrant children in Japan: A comparison across children s nationalities, Asian and Pacific Migration Journal, 17(2): Dwyer, J.W., and P. Lovell 1990 Earnings differentials between whites and Japanese: The case of Brazil, Sociological Perspectives, 33(2): Feliciano, C. and R.G. Rumbaut 2005 Gendered paths: Educational and occupational expectations and outcomes among adult children of immigrants, Ethnic and Racial Studies, 28(6): Fukuoka, Y Lives of Young Koreans in Japan, Trans Pacific Press, Melbourne. Genda, Y A Nagging Sense of Job Insecurity: The New Reality Facing Japanese Youth, International House of Japan, Tokyo, Japan.