Social Capital and Immigrant Women s Employment in the State of Hawaii

Transcription

1 Graduate Journal of Asia-Pacific Studies 4:1 (2006) Social Capital and Immigrant Women s Employment in the State of Hawaii Yasuko URANO University of Hawaii at Manoa, United States of America THE PERSONAL Responsibility and Work Opportunity Reconciliation Act (PRWORA), signed into law in 1996 by the U.S. president Bill Clinton, pressures non-u.s. citizens to become economically independent by restricting their access to public welfare benefits. 1 For example, during the first five years of the PRWORA, new immigrants were not eligible for means-tested federal programs, including Temporary Assistance for Needy Families, Medicaid, Supplemental Security Income, Food Stamps, and the Children s Health Insurance Program. 2 In the void created by the reform, the State of Hawaii decided to use its own funds to partially support immigrants 3 although many non-citizens also fail to qualify for these state funded programs. 4 Even with this limited support, the State of Hawaii remains desirable for many immigrants and its population continues to grow. 5 In order to live without public support, many new immigrants seek employment; this is defined as either being (1) with a job and at work as paid employees, working in their own business or profession, working on their own farm, or working 15 hours or more as unpaid workers on a family farm or in a family business ; or (2) with a job but not at work those who have jobs or businesses from which they are temporarily absent due to industrial dispute, or personal reasons. 6 Among first-generation Asian-Pacific immigrant women aged 18 to 64, employment rates range from 47.8 per cent to 68.7 per cent, depending on the ethnic group. 7 Scholars have often analysed such diversified employment rates from the standpoints of human capital (e.g. educational attainment), 8 rational choice (i.e. decision to work according to family s situation) 9 and ethnicity (e.g. cultural norms of female employment). 10 Recently however, attention has turned to the role of social capital. 11 Social capital, as defined by Coleman, is the relationships among family and the social relationships that exist between family and the institutions of the community. 12 This study aims to examine the relationship between social capital and women s employment, with a focus on first-generation immigrant women. Following recent welfare reforms, there has been more serious scholarly attention paid to immigrants employment; 13 however, few studies have looked specifically at immigrant women and whether they choose to obtain employment using social capital, or if they prefer domestic activity. This study considers these issues as well as exploring the influence of ethnic and cultural diversity on the nature of employment. This is accomplished by focusing on the State of Hawaii, as it provides an effective case study of Asian-Pacific Urano/Social Capital and Immigrant Women s Employment 65

2 female migration. In this research, first-generation immigrant women with Chinese, Filipino, Japanese, Korean, Vietnamese and Pacific Island backgrounds were studied because they represent the major immigrant groups in the State of Hawaii. 14 Theoretical Backgrounds In this study, four distinct theoretical foci are employed to investigate the factors that contribute to the differences between immigrant women s employment rates. These foci are human capital, ethnicity, choice and social capital. Human capital theory explains the relationship between human capital investment and wages, salaries and employment. 15 Three important aspects of human capital are (1) English proficiency a prerequisite for working, 16 (2) educational attainment an employment-based H1b visa requires minimally a U.S. bachelor's degree, or the equivalent, 17 and (3) U.S. citizenship which is important and valuable because of benefits such as increased federal job opportunities and entitlement to important public benefits. However, citizenship is not easy to obtain. In general, immigrants become eligible for citizenship after obtaining permanent residency (a green card ) and living as a good citizen for five years, or after being married to a U.S. citizen for at least three years without legal separation. 18 Eligibility also requires good moral character, a decent level of English ability and sufficient knowledge of American history and government. 19 Ethnicity is another important aspect of individual characteristics. Cultural theory explains that ethnicity influences the cultural norms that dictate lifestyle and the family system. 20 Within cultural theory, there are several key debates around issues relating to immigrants. The cultural assimilation theory, for example, suggests that immigrants are gradually assimilated into mainstream society. In contrast, cultural pluralism argues that immigrants are more likely to maintain their cultural traditions, even in the face of assimilating pressures. 21 Finally, employment research recognizes that first-generation immigrants present unique employment patterns which are similar to those in their native region/country. 22 Consequently, it is pointed out that married East Asian women prefer domestic activities to taking paid work, whereas Filipino women choose to work regardless of their marital status. 23 From another theoretical perspective, the choice theory of family structure suggests that women are encouraged to decide whether to work by rationally considering the situation in which they find themselves. Previous studies have revealed that the significant factors affecting a woman s employment include marital status, 24 the presence of young children, 25 and the family income without the woman s income. 26 Those who are married and have young children may prefer to use their time domestically (taking care of the home and family members) rather than working. Other scholars point out that, where family livelihood depends on a woman s income, women obtain employment and remain employed, regardless of marital status or the presence of young children. 27 Finally, there has been much scholarly attention paid to the positive impact of social capital on employment status. These studies have defined some aspects of social 66

3 relations as social capital and refer to the previously cited work of Coleman. In the context of child development, Coleman refers to social capital of the family and the community and defines the former as the relation between children and parents (and, when families include other members, relationships with them as well). 28 Additionally, Coleman states that the latter resides in the functional community, the actual social relationships that exist between parents, in the closure exhibited by the structure of relations, and in the parents relations with the institutions of the community. 29 This definition partly overlaps Putnam s definition of social capital. Putnam considers social capital as features of social organization, such as trust, norms and networks, 30 and believes that community involvement is an important aspect of social capital. 31 These ideas suggest that social capital plays an important role in immigrant women s employment. For example, members of a small family who recently migrated may not have well-developed social capital since they have little to help each other with and are not familiar with their neighbours and the wider community. In contrast, women who migrated earlier and live with subfamily defined as either (1) a single parent who lives with at least one child under 18 years of age or (2) a married couple, with or without children may obtain social capital from them. Research Question My primary research question aims to explore how first-generation immigrant women of working age (18 to 64) in the State of Hawaii take advantage of their potential resources, in the form of social capital, to gain employment. It is important to focus on employment outcomes because participation in the labor force, which is defined as (1) working or (2) looking for work, 32 does not assure acquisition of employment; some people may not get a job because they lack human or social capital. Data and Statistical Models This research uses a sample obtained from the Census 2000, 5-Percent Public Use Microdata Sample Files (PUMS) for the State of Hawaii. 33 I use SPSS and conduct binary logistic regression. Binary logistic regression is the most appropriate form of analysis since: (1) the dependent variable is dichotomous; 34 (2) it makes it possible to assess the percentage of variance in the dependent variable explained by the independent variables ; 35 and (3) interaction effects can be examined. 36 The regression equation can be written as: Logit(π) = α+β₁x₁ +βixi Here x = independent variables, α = the constant of the equation and β = the beta coefficient (B) of the independent variables. In order to run the logistic regression, I constructed statistical models based on the theoretical backgrounds and coded variables. When I created models, I first examined effects that were clearly revealed by previous studies. After controlling conventional variables, I explored the impacts of social capital, since they had yet to be investigated. Urano/Social Capital and Immigrant Women s Employment 67

4 Finally, I introduced the interaction variable between presence of children under six years of age and presence of subfamily. The first model controls only for individual characteristic variables to investigate the impact of those characteristics on employment, as shown in Figure 1. Individual characteristics include age, ethnicity, educational attainment, English ability and citizenship, as Figure 1 illustrates. I coded these variables as shown in Table 1. Age is expected to have a quadratic effect on employment (i.e. women tend to work up to a certain age, then they tend not to work later). 37 In the current study, two variables were included: a continuous variable and an age-squared variable. I used B values of these variables to calculate the maximum point of age. The maximum point can be calculated as -B of age / 2*B of age-squared. 38 Individual Characteristics X1 Age X2 Age-Squared X3-9 Ethnicity X10-11 Educational Attainment X12 English Proficiency X13 Citizenship Status Employed Unemployed Figure 1: The first model For analysing the impact of ethnicity, seven dummy variables were created: firstgeneration Chinese, Filipino, Japanese, Korean, Vietnamese, Pacific Islander and other immigrants. Native-born American women were left out to prevent multicollinearity, and used as the reference category. They include those who were (1) born in the United States, Puerto Rico, or U.S. Island Areas 39 or, (2) born abroad of American parent(s). 40 In all groups, those fitting into only one ethnic category are included. Pacific Islander consists of all non native-born women having origins in any of the original peoples of Hawaii, Guam, Samoa or other Pacific Islands. It includes people who indicate their race as native Hawaiian, Guamanian or Chamorro, Samoan and other Pacific Islander. 41 Other immigrants consist of Caucasian, African, other Asian, some other race and a mixture of two or more major races. Educational level was divided into three groups: without high school diploma, high school graduate or some college and college graduate or studies beyond college graduation. In coding the variable of English proficiency, individuals who have proficiency in English were defined as those who reported they could speak English very well. Finally, citizenship status was divided into (1) native-born American and naturalized immigrant and (2) non-u.s. citizen. The regression equation for the first model can be written as: Logit(π) = α+β₁x₁+ +β₁₃x₁₃ 68

5 Using B values, a researcher can calculate by hand and identify the probability of a specific sample. For example, using the result in Table 4, the probability of an employed Chinese woman, aged 25, college graduate, speaking English very well and non-u.s. citizen, can be written as: Logit (p) = (25)-0.002(625)+0.276(1)+0.551(1)+0.149(1) = exp⁰ ⁵⁵⁷ = (odds) exp⁰ ⁵⁵⁷/1+ exp⁰ ⁵⁵⁷ = (probability) Variables Employment Status Age Ethnicity Educational Attainment English Ability Citizenship Marital Status Number of Family Members Presence of Children < 6 Number of Children < 18 Other Family Members Income Presence of Subfamily Length of Stay in the U.S. Interaction Table 1: List of Variables and the Measures Coding of Variables Employed=1, Unemployed=0. (X1) coded as continuous (X2) coded as squared values (X3) Chinese=1, others=0; (X4) Filipino=1, others=0; (X5) Japanese=1, others=0; (X6) Korean=1, others=0; (X7) Vietnamese=1, others=0; (X8) Pacific Islanders=1, others=0; (X9) Other immigrants=1, others=0. Native born women were left out to use as a reference category. (X10) High School Graduate or Some College Education=1, others=0, (X11) College Graduate or Beyond College Education =1, others=0. Those without high school diploma were left out to use as a reference category. (X12) Speaking English very well=1, others=0. (X13) U.S. citizen=1, non-u.s. citizen=0. (X14) Separated=1, others=0; (X15) Divorced=1, others=0; (X16) Widowed=1, others=0; (X17) Never-married=1, others=0. Married status was left out to use as a reference category. (X18) Coded as Continuous. (X19) Yes=1, No=0. (X20) Coded as Continuous. (X21) $5000 to $9999=1, others=0, (X22) $10000 to $14999=1, others=0, (X23) $15000 to $24999=1, others=0, (X24) $25000 to $34999=1, others=0, (X25) $35000 to $49999=1, others=0, (X26) $50000 to $74999=1, others=0, (X27) $75000 to $99999=1, others=0, (X28) $ and over=1, others=0. Having less than $5000 was left out to use as a reference category. (X29) Yes=1, No=0. (X30) More than 5 years=1, less than 5 years=0. (X31) Living with children under 6 and subfamily=1, others=0. Urano/Social Capital and Immigrant Women s Employment 69

6 The second model introduces family characteristic variables after controlling for individual characteristics, as shown in Figure 2. Family characteristics include marital status, presence of children under six years of age, number of children under 18 years of age, number of family members and total family income not including a respondent s income. I coded these variables as shown in Table 1. Marital status was divided into five categories: married, separated, divorced, widowed and never-married. I created a dummy variable for the presence of children under six years of age and coded a continuous variable as the number of children under 18 years of age. I added the variable for the number of family members because a big family increases both the need for domestic activity and a greater family income, regardless of the presence of young children. The final family characteristics variable focuses on the economic situation. I created categorical variables using the U.S. Census income clusters 42 under $5000, $5000 to $9999, $10000 to $14999, $15000 to $24999, $25000 to $34999, $35000 to $49999, $50000 to $74999, $75000 to $99999 and more than $ Those with under $5000 were used as a reference category. Individual Characteristics X1-X13 Family Characteristics X14-17 Marital Status X18 Number of Family Members X19 Presence of Children <6 X20 Number of Children < 18 X21-28 Total Family Income without a Respondent s Income Employed Unemployed Figure 2: The second model The regression equation for the second model can be written as: Logit(π) = α+β₁x₁+ +β₂₈x₂₈ The probability of a specific sample can be obtained by using the equation. For example, using results from Table 4, the probability of an employed Japanese woman aged 29, college graduate, speaking English very well, a non-u.s. citizen, never-married, living alone, can be written as follows: Logit (p)= (29)-0.003(29*29) (1)+0.545(1)+0.124(1)+0.279(1)+0.005(1)=0.227 exp⁰ ²²⁷=1.255 (odds) exp⁰ ²²⁷/1+ exp⁰ ²²⁷=0.557 (probability) 70

7 The third model controls for social capital to investigate the impact on employment after controlling for individual and family characteristics, as shown in Figure 3. The social capital variables are presence of subfamily in the household and length of stay in the United States. I created a dummy variable for the presence of subfamily in the household, as shown in Table 1. Previous studies identify those who arrived in the United States within the past five years as the most recent immigrants, and thus socially and culturally unfamiliar with American society. 43 These immigrants are unlikely to have well-developed social capital in a community. In this study, I employed a period of five years and created a dummy variable for length of stay in the United States. Individual Characteristics X1-X13 Family Characteristics X14-X28 Social Capital Characteristics X29 Presence of Subfamily X30 Length of Stay in the United States Employed Unemployed Figure 3: The third model The equation for the model can be written as: Logit(π) = α+β₁x₁ +β₃₀x₃₀ When it comes to the probability of a specific sample, for example, the probability of employment for a Filipino woman, aged 31, high school graduate, speaking English very well, a non-u.s. citizen, married, living with her husband who obtains income $30000, has a child aged three and a subfamily, having lived in the United States for ten years, can be written as follows: Logit (p)= (31)-0.003(31*31)+0.427(1)+0.584(1)+0.112(1) (3)-0.414(1)-0.098(1)+0.265(1)-0.201(1)+0.310(1)=0.226 exp⁰. ²²⁶=1.254 (odds) exp⁰. ²²⁶/1+ exp⁰. ²²⁶=0.556 (probability) The final model controls for interaction effect to investigate the impact of interaction between presence of children under six years of age and presence of Urano/Social Capital and Immigrant Women s Employment 71

8 subfamily in the household, after controlling for individual, family and social capital variables, as shown in Figure 4. The equation for the model can be written as: Logit(π) = α+β₁x₁ +β₃₁x₃₁ To demonstrate the probability of a specific sample, using an employed Korean woman aged 33, college graduate, speaking English very well, non-u.s. citizen, married, lives with her husband who obtains income $45000, has a child aged two and a subfamily, and has lived in the United States for two years, can be written as follows: Logit (p)= (33)-0.003(33*33)-0.411(1)+0.550(1)+0.111(1) (3)-0.447(1)-0.100(1)+0.003(1)-0.242(1)+0.588(1)= exp ⁰ ⁵⁹²=0.553 (odds) exp ⁰ ⁵⁹²/1+ exp ⁰ ⁵⁹²=0.356 (probability) Individual Characteristics X1-X13 Family Characteristics X14-X28 Interaction X31 (X19*X29) Social Capital Characteristics X29-X30 Figure 4: The fourth model Employed Unemployed For the analysis, it is necessary to calculate by hand and interpret the main and interaction effects according to the research interests. My research interest is how the impact of social capital of the subfamily varies depending on the presence of young children. In order to identify the impact, I multiplied the odds of the presence of subfamily by the odds of interaction. Using SPSS, I first looked at frequency tables and analysed descriptive statistics looking at the percentage of the total frequency and means of each variable among each group. After I ran the logistic regression, I looked at (1) the chi-square goodness-of-fit test, (2) the Nagelkerke and the Cox and Snell R-square values and (3) information on variables in the equation. In the findings, I first assess the fit of the overall model by using the chi-square goodness-of-fit test. If the test is significant, this means that the constructed model fits the data. Second, I assess the fit of the model of the data by looking at R-square values. A greater R-square value is better. 44 For example, if the R-square value of Model 2 is greater than that of Model 1, it indicates that Model 2 fits better than Model 1. Third, I 72

9 analysed the impact size of each variable and interpreted it using the odds ratio. In the binary logistic regression model, SPSS provides the odds ratio of being in the higher group compared to the lower group or the reference group. 45 For example, if the odds ratio of U.S. citizenship status is 1.30, this indicates that having U.S. citizenship makes an individual 30 per cent more likely to be employed than those without U.S. citizenship; if the odds ratio of being Japanese is 0.95, this indicates that Japanese immigrant women are 5 per cent less likely to be employed than native-born American women. I use B values specifically when I interpret the impact of age as noted earlier. Findings Table 2 shows the characteristics of native-born American and first-generation immigrant women. In terms of employment status, Filipinos had the highest rate (68.7%) and native-born American women had the second highest (67.7%), while Pacific Islanders had the lowest rate (47.8%). Employment rates of the other ethnicities were: Chinese (66.4%), Japanese (59.9%), Korean (53.4%) and Vietnamese (54.1%). The Japanese were the oldest (43.4 years on average), while both Vietnamese and Pacific Islanders were the youngest (37.1 years on average). When it comes to human capital characteristics, Vietnamese women had the highest rate of those without high school diplomas (39.3%), whereas native-born American women had the lowest rate (7.2%). Native-born American women had the highest rate of having high school diplomas or some college education (66.6%), whereas Chinese women had the lowest rate (43.0%). The highest rate of college graduates or those having more than a college education was Japanese women (27.4%), whereas the lowest was Pacific Island women (7.5%). English proficiency ranged from 70.5 per cent to 33.6 per cent across the ethnic groups, with the other immigrant women group having the highest rate and Vietnamese women having the lowest rate. The other rates were: Chinese (39.7%), Filipino (49.9%), Japanese (41.3%), Korean (36.6%) and Pacific Islanders (61.7%). Acquisition of citizenship was also differentiated by ethnic groups: Chinese (68.6%), Filipino (67.7%), Japanese (33.1%), Korean (65.1%), Vietnamese (71.3%) and Pacific Islanders (48.3%). Among the other immigrants group, 47.5 per cent were naturalized U.S. citizens. From the standpoint of marital status, native-born Americans had the lowest rate of married status (54.1%), whereas Filipinos had the highest rate (70.5%). Koreans had the highest rate of separation (3.1%), whereas Pacific Islanders had the lowest rate (0.5%). Koreans had the highest rate of divorced status (14.8%), whereas Filipinos had the lowest rate (4.7%). Native-born Americans had the lowest rate of being widowed (2.6%), whereas Filipinos had the highest rate (6.8%). Pacific Islanders had the highest rate of never being married (30.8%), whereas Filipinos had the lowest rate (16.1%). Pacific Islanders had the highest rate of the presence of children under six years of age (23.4%), whereas Japanese had the lowest rate (10.2%). Pacific Islanders had the most children under 18 years of age (1.2 on average), whereas Japanese had the fewest (0.5 on average). Urano/Social Capital and Immigrant Women s Employment 73

10 Variables Native-born American Chinese Filipino Japanese Korean Vietnamese Pacific Islander Other Employment Per cent(sd) 67.7(0.47) Per cent(sd) 66.4(0.47) Per cent(sd) 68.7(0.46) Per cent(sd) 59.9(0.49) Per cent(sd) 53.4(0.50) Per cent(sd) 54.1(0.50) Per cent(sd) 47.8(0.50) Per cent(sd) 54.9(0.50) Age (mean) < High School High School Graduate College Graduate English Proficiency Citizenship 39.7(12.6) 7.2(0.51) 66.6(0.17) 26.2(0.00) 97.0(0.15) (12.27) 30.3(0.46) 43.0(0.50) 26.7(0.89) 39.7(0.49) 68.6(0.46) 41.9(12.27) 26.3(0.44) 54.5(0.50) 19.2(0.79) 49.9(0.50) 67.7(0.47) 43.4(12.57) 10.2(0.30) 62.4(0.49) 27.4(0.89) 41.3(0.49) 33.1(0.47) 42.5(11.40) 22.4(0.42) 61.1(0.49) 16.5(0.74) 36.6(0.48) 65.1(0.48) 39.10(11.7) 39.3(0.49) 47.5(0.50) 13.1(0.68) 33.6(0.47) 71.3(0.45) 37.1(11.78) 32.8(0.47) 59.7(0.49) 7.5(0.53) 61.7(0.49) 43.8(0.50) 40.2(12.51) 16.3(0.37) 56.6(0.50) 27.1(0.89) 70.5(0.46) 47.5(0.50) Married Separated Divorced Widowed Never-Married Number of Family Members(mean) Presence of Children < 6 Number of Children < 18(mean) 54.1(0.50) 2.1(0.14) 12.3(0.33) 2.6(0.16) 28.9(0.45) 3.2(2.17) 16.2(0.36) 0.8(1.12) 67.4(0.47) 0.7(0.08) 6.1(0.24) 4.5(0.21) 21.3(0.41) 3.2(1.97) 15.6(0.36) 0.6(0.94) 70.5(0.46) 1.9(0.14) 4.7(0.21) 6.8(0.25) 16.1(0.37) 5.0(2.86) 17.8(0.38) 0.9(1.15) 63.7(0.48) 2.0(0.14) 10.2(0.30) 5.7(0.23) 18.4(0.39) 2.3(1.68) 10.2(0.30) 0.5(0.80) 61.3(0.49) 3.1(0.17) 14.8(0.36) 3.8(0.19) 17.0(0.38) 2.6(1.72) 10.7(0.31) 0.6(0.88) 59.0(0.49) 1.6(0.13) 10.7(0.31) 4.1(0.20) 24.6(0.43) 3.2(1.83) 19.7(0.40) 0.89(1.09) 59.2(0.49) 0.5(0.70) 5.0(0.22) 4.5(0.21) 30.8(0.46) 5.1(2.92) 23.4(0.42) 1.2(1.46) 61.3(0.48) 2.3(0.15) 12.5(0.33) 4.8(0.21) 19.0(0.39) 3.0(2.15) 19.0(0.39) 0.8(1.10) Other Family Members Income < $5000 $5000 to 9999 $10000 to $15000 to $25000 to $35000 to $50000 to $75000 to $ (0.48) 18.2(0.39) 15.8(0.36) 18.3(0.39) 6.5(0.25) 3.1(0.17) 1.6(0.12) 0.6(0.08) 0.7(0.08) 34.3(0.48) 23.2(0.42) 15.6(0.36) 17.7(0.38) 4.0(0.20) 3.1(0.17) 1.9(0.14) 0.2(0.05) 25.2(0.43) 33.5(0.47) 22.4(0.42) 14.5(0.35) 3.1(0.17) 0.8(0.09) 0.5(0.07) 34.9(0.48) 14.3(0.35) 14.5(0.35) 19.5(0.40) 8.6(0.28) 6.1(0.24) 0.9(0.09) 0.9(0.09) 0.2(0.05) 38.9(0.49) 13.7(0.34) 14.5(0.35) 19.8(0.40) 6.4(0.24) 3.3(0.18) 2.0(0.14) 1.3(0.11) 47.5(0.50) 23.8(0.43) 12.3(0.33) 9.0(0.29) 4.1(0.20) 2.5(0.16) 0.8(0.09) 49.3(0.50) 23.8(0.43) 12.9(0.34) 9.5(0.29) 2.0(0.14) 1.0(0.10) 0.5(0.07) 39.5(0.49) 20.0(0.40) 13.3(0.34) 15.7(0.36) 5.7(0.23) 2.5(0.16) 1.8(0.13) 0.6(0.78) 1.0(0.10) Presence of Subfamily Recent Immigration (reference) Children<6 X Subfamily N 12.5(0.33) 0.4(0.64) (0.33) 15.1(0.36) 0.9(0.10) 423 Table 2: Percent, Means and Standard Deviation of the Variables 28.7(0.45) 15.8(0.37) 3.6(0.19) (0.20) 26.1(0.44) (0.24) 10.4(0.31) 0.3(0.50) (0.25) 14.8(0.36) 0.8(0.09) (0.45) 22.9(0.42) 2.5(0.16) (0.26) 19.5(0.40) 0.5(0.69)

11 When it comes to family size, the mean number of family members among nativeborn Americans was 3.2. Pacific Islanders had the largest families (5.1) and Filipinos had the second largest (5.0). In contrast, Japanese had the smallest families (2.3). Total family annual income not including a respondent s income also varied across ethnic groups. The distribution of income among native-born Americans was: those with less than $5000 (35.1%), $5000 to $9999 (18.2%), $10000 to $14999 (15.8%), $15000 to $24999 (18.3%), $25000 to $34999 (6.5%), $35000 to $49999 (3.1%), $50000 to $74999 (1.6%), $75000 to $99999 (0.6%) and more than $ (0.7%).Although Pacific Islanders had more than five family members on average, almost half of them (49.3%) lived on less than $5000, not including the respondent s income. Since 52.2 per cent of women were not engaged in paid work, many were assumed to live on a very small budget. In contrast, Filipinos had bigger families and only 25.2 per cent of them lived on less than $5000, not including the respondent s income. With respect to social capital, the presence of subfamily among native-born Americans was 12.5 per cent. Filipinos had the highest rate (28.7%) and Pacific Islanders, the second highest rate (27.4%). In contrast, Japanese had the lowest rate (4.3%). The rate of recent immigration also varied across groups. Japanese were the highest (26.1%), whereas Koreans were the lowest (10.4%). Among the other groups: Chinese (15.1%), Filipino (15.8%), Vietnamese (14.8%), Pacific Islander (22.9%) and other immigrants (19.5%), were recent immigrants. Regarding women having children under six years of age and living with a subfamily, Filipinos had the highest rate (3.6%) Pacific Islanders had the second highest rate (2.5%), while Japanese had 0 per cent. Table 3 illustrates Models 1 and 2, showing the statistical impact of independent variables. The first model was developed to investigate how individual characteristics are associated with female employment and are statistically significant (χ²= , p<0.001 with df=13). The Nagelkerke R-square suggests that 9.3 per cent of the variance of the dependent variable is explained by the independent variables (individual characteristics, in this model), whereas the Cox and Snell R-square is 6.7 per cent. Age was observed to have a highly significant effect on the likelihood of employment. The effect of age was curvilinear, with the odds of employment increasing with age up to about years and then declining. As expected, the likelihood of employment increased with the level of educational attainment. Those with a high school education or some college education were 80 per cent more likely than the reference group (those without high school diploma) to obtain employment. Those who had graduated from college or with studies beyond college graduation were 74 per cent more likely to be employed than the reference group. Those who spoke English well were 16 per cent more likely to be employed than those who did not speak English well. Having U.S. citizenship was also observed to be significant: women who had U.S. citizenship were 44 per cent more likely to have a job than non-u.s. citizens. Being first-generation immigrants also proved to be a significant determinant of employment. However, the impact varied across ethnic groups. When compared to the reference group (native-born American women), Chinese women were 32 per cent more Urano/Social Capital and Immigrant Women s Employment 75

12 likely and Filipino women were 47 per cent more likely to obtain employment. On the other hand, Koreans were 30 per cent less likely, Pacific Islanders were 33 per cent less likely and other immigrants were 26 per cent less likely to obtain employment than the reference group. The odds of employment for Japanese and Vietnamese were not found to be significantly different from the reference group. Model 2 included family circumstance variables to explore whether or not women considered their family situation in obtaining employment. This model was statistically significant (χ²= , p<0.001 with df=28). The Nagelkerke R-square suggests that 11.5 per cent of the variance in the dependent variable is explained by the independent variables, whereas the Cox and Snell R-square was 8.3 per cent. Comparing R-square values for the first and the second models, the second model provides a slightly better fit for the data. The results show that women who were divorced or never-married were more likely to obtain employment than the reference group (married women); divorced women were 41 percent more likely to be employed than the reference group while never-married women were 32 percent more likely to be employed than the reference group. Regarding the presence of children, it was observed that the presence of children under age six and the number of children under age 18 made women less likely to be employed. Women with children under six years of age were 34 per cent less likely to be employed than those who did not have children under age six. For women with children under age 18, the odds of obtaining employment decrease 6 per cent for each additional child. Total family annual income, not including the respondent s income, was shown to be significant, with the likelihood of women s employment varying with family income levels. When compared to the reference group (those having less than $5000 income), those having $5000 to $9999 were 11 per cent more likely to be employed, those having $10,000 to $14,999 per year were 36 per cent more likely to be employed, those having $15,000 to $24,999 were 38 per cent more likely to be employed and those having $25,000 to $34,999 were 34 per cent more likely to be employed. Having $100,000 and over was observed to have a negative impact on the likelihood of employment. When other family members obtained more than $100,000 per year, women were 48 per cent less likely to be employed than the reference group. The number of family members was not shown to be significant. Even after introducing family characteristic variables, individual characteristic variables were shown to have a similar impact. Only the maximum point of age was shown to be younger. The odds of employment increased with age up to about years and then declined. 76

13 Independent Variables Model 1 Model 2 Model 3 Model 4 Age Age-squared Native-born American First-Generation Chinese Filipino Japanese Korean Vietnamese Pacific Islander Other <High Shool High School Graduate College Graduate English Proficiency Citizenship Married Separated Divorced Widowed Never-Married Number of Family Members Presence of Children < 6 Number of Children < 18 Other Family Members Income < 5000 $5000 to $9999 $10000 to $14999 $15000 to $24999 $25000 to $34999 $35000 to $49999 $50000 to $74999 $75000 to $99999 $ Presence of Subfamily Length of Stay in the United States >5 Children under 6 x Presence of Subfamily Intercept Nagelkerke R-square (Cox & Snell R-square) Chi-square -2 Log likelihood Degrees of Freedom Odds(B Coefficient) 1.208***(0.189) 0.998***(-0.002) 1.317*(0.276) 1.472***(0.387) 0.988(-0.012) 0.704**(-0.351) 0.814(-0.206) 0.673*(-0.395) 0.739***(-0.303) 1.796***(0.586) 1.735***(0.551) 1.160*(0.149) 1.442***(0.366) 0.020***(-3.894) 0.093(0.067) *** Odds(B Coefficient) 1.247***(0.221) 0.997***(-0.003) 1.329*(0.285) 1.516***(0.416) 0.951(-0.050) 0.675**(-0.393) 0.854(-0.158) 0.715*(-0.335) 0.762**(-0.272) 1.781***(0.577) 1.725***(0.545) 1.132*(0.124) 1.409***(0.343) 1.157(0.146) 1.409***(0.343) 0.912(-0.092) 1.321***(0.279) 1.005(0.005) 0.661***(-0.413) 0.939**(-0.063) 1.113*(0.107) 1.359***(0.306) 1.383***(0.324) 1.343***(0.295) 1.045(0.044) 0.857(-0.155) 0.820(-0.199) 0.523***(-0.649) 0.010***(-4.562) 0.115(0.083) *** Odds(B Coefficient) 1.247***(0.221) 0.997***(-0.003) 1.340*(0.293) 1.532***(0.427) 0.957(-0.044) 0.666***(-0.407) 0.852(-0.160) 0.717*(-0.333) 0.760**(-0.274) 1.793***(0584) 1.733***(0.550) 1.119(0.112) 1.273**(0.241) 1.163(0.151) 1.419***(0.350) 0.921(-0.083) 1.314***(0.273) 1.030**(0.029) 0.661***(-0.414) 0.907***(-0.098) 1.098(0.094) 1.330***(0.285) 1.343***(0.295) 1.303**(0.265) 1.009(0.009) 0.833(-0.182) 0.801(-0.222) 0.506***(-0.682) 0.818**(-0.201) 1.364**(0.310) 0.009***(-4.756) 0.117(0.084) *** Odds(B Coefficient) 1.248***(0.221) 0.997***(-0.003) 1.336*(0.289) 1.513***(0.414) 0.956(-0.045) 0.663***(-0.411) 0.849(-0.164) 0.715*(-0.335) 0.760**(-0.274) 1.795***(0.585) 1.734***(0.550) 1.117(0.111) 1.724**(0.242) 1.162(0.150) 1.416***(0.348) 0.925(-0.078) 1.308***(0.269) 1.030*(0.030) 0.640***(-0.447) 0.905***(-0.100) 1.098(0.094) 1.333***(0.288) 1.344***(0.295) 1.298**(0.261) 1.003(0.003) 0.832(-0.184) 0.800(-0.223) 0.502***(-0.688) 0.785***(-0.242) 1.371***(0.316) 1.800**(0.588) 0.009***(-4.760) 0.117(0.085) *** Table 3: Odds Ratios and B Coefficient for Female Employment Equation (*p<0.05, **p<0.01, ***p<0.001) Urano/Social Capital and Immigrant Women s Employment 77

14 The third model includes social capital characteristics and their impact on the likelihood of employment (Table 4). This model was statistically significant (χ²= , p<0.001 with df=30). The Nagelkerke R-square suggests that 11.7 per cent of the variance of the dependent variable is explained by the independent variables. The Cox and Snell R- square indicates 8.4 per cent. In comparing R-square values for the second and third models, the third model is not a big improvement over the second model. Living with subfamily was shown to have a negative impact on employment. Women with subfamily were 18 per cent less likely to obtain employment than those without subfamily. Length of stay in the United States was shown to be significant in Model 3. Those who had been in the United States for more than five years were 36 per cent more likely to obtain employment than those with less than five years residency. Introducing social capital variables affected other variables. Speaking English well was no longer significant. Having $5000 to $9999 income was also no longer significant. On the other hand, the impact of the number in family became significant: the odds of obtaining employment increased by 3 per cent for each additional family member. The final model was developed to investigate the interaction effect between the presence of children under age six and the presence of subfamily (Table 4). This model was statistically significant (χ²= , p<0.001 with df=31). The Nagelkerke R-square suggests that 11.7 per cent of the variance of the dependent variable is explained by the independent variables. The Cox and Snell R-square indicates 8.5 per cent. The R-square values for the final model were not improved when compared to the third model. However, since the interaction variable was shown to be strongly significant, it should be included in the model. After introducing interaction variables, individual characteristic variables were basically shown to have a similar impact when the models were compared. The interaction variable was shown to be significant and the odds ratio was I multiplied the odds of the presence of subfamily by the odds of interaction: 0.79*1.80 = This means that those who had young children and lived with subfamily were 42 per cent more likely to obtain employment than those who had young children but did not live with subfamily. Discussion and Conclusion These findings support human capital theory by suggesting that women who have an appropriate educational background and citizenship have an increased likelihood of employment. However, the result showing that English proficiency is not significant is not consistent with previous studies. This may be a result of using the State of Hawaii and its census data, and reflect the multicultural nature of its population and employment opportunities. As expected from choice theory, family characteristics also provide a strong explanation for choosing not to obtain employment. Women appear to prefer domestic activity to working if they have young children, or if other family members earn sufficient money. The results also highlight diversified employment patterns depending on ethnic groups. My findings show that, with the exception of Filipino and Chinese women, firstgeneration immigrant women are less likely than native-born Americans to be 78

15 employed. This suggests some potential guidelines for a detailed analysis of cultural values of female employment. Some aspects of social capital of family can provide further explanation for the likelihood of employment. Living with subfamily decreased the likelihood of employment in Models 1, 2 and 3. A possible explanation is that some women chose not to work as the subfamily earned sufficient money, and they intended to contribute by helping in the home. In terms of social capital, such women may not choose to use social capital to obtain employment; instead, they may choose to supply social capital within their subfamily. However, the final model indicates an additional attribute of women who had young children and lived with subfamily. Such women were more likely to work than women with young children but without subfamily. Since taking care of young children requires both time and money, women living with subfamily may have support and assistance (e.g. baby-sitting) which enables them to work. Regarding the social capital of community, recent immigration decreased the likelihood of employment. A possible explanation is that many recent immigrant women entered the United States as non-working spouses they instead chose to supply social capital to their working family members. Another explanation is that such women quickly become accustomed to living in the new environment and have built social capital. Consequently, even non-working, recent immigrant women, who have lived in the United States more than five years, were more likely to gain employment and enjoy both domestic activity and working. Finally, three implications for future research are pointed out. First, other social capital elements, such as closeness with the subfamily, personal networks and affiliation with ethnic or community groups may also be important factors affecting the decision to obtain employment. Second, more detailed variables on ethnicity-oriented social capital may be able to answer further research questions regarding how the impact of social capital on employment varies across ethnic groups. Third, more comparative macro analysis at the interstate (e.g. Hawaii vs. California) and international scale is warranted. Such analyses are likely to provide further evidence of the dynamic trends of Asian- Pacific migration in the United States as well as other countries where Asian-Pacific migration is significant. In conclusion, the overall findings tell many stories about social capital and employment. First, the results indicate that for women who have young children, subfamily is an important social capital for gaining employment, but this resource does not have the same impact among all women. Second, although welfare reform keeps immigrants from using public benefits, recent immigrant women are less likely to work. Either they intend to become a non-working spouse and provide social capital to their working family members, or they make great efforts to build up social capital soon after arriving in the United Sates and are more likely to be working after the initial five-year period. Because of this dichotomy, employment research should pay more attention to whether immigrant women choose to use or supply social capital. These findings open new perspectives on social capital and for studying Asian-Pacific female migration and employment trends. Urano/Social Capital and Immigrant Women s Employment 79

16 NOTES 1 The United States General Accounting Office, Welfare Reform, online, July 1998, available at: (11 April 2006) Vik Jolly, Where is the welcome mat?, online, 27 September 1996, available at: (11 April 2006). 5 The United States Census Bureau FactFinder, online, nd, available at: (8 January 2006). 6 The U.S. Census 2000 PUMS, Glossary, online, nd, p. 6(E), available at: (8 January 2006). 7 U.S. Census PUMS. 8 Haya Stier, Immigrant Women Go to Work: Analysis of Immigrant Wives Labor Supply for Six Asian Group, Social Science Quarterly (SSQ), 72, 1 (1991), pp Terrence Casey and Kevin Christ, Social Capital and Economic Performance in the American States, SSQ, 86, 4 (2005), pp James S. Coleman, The Creation and Destruction of Social Capital: Implications for the Law, Journal of Law, Ethics, and Public Policy (JLEPP), 3 (1988), pp George J. Borjas, Welfare Reform, Labor Supply, and Health Insurance in the Immigrant Population, Journal of Health Economics, 22 (2003), pp U.S. Census PUMS. 15 Robert. F. Shoeni, Labor Market Outcomes of Immigrant Women in the United States: 1970 to 1990, International Migration Review, 32, 1 (1998), pp U.S. Citizenship and Immigration Services, online, nd, available at: (15 April 2006) Monica McGoldrick, Joe Giordano and John K. Pearce, eds, Ethnicity and Family Therapy, 3rd ed., Guildford, New York, 1996, pp Bill Ong Hing, To be an American cultural pluralism and the rhetoric of assimilation, New York University Press, New York, 1997, pp Heather Antecol and Kelly Bedard, The Decision to Work by Married Immigrant Women: The Role of Extended Family Households, Claremont Colleges working papers in economics, 34, online, October 2002, available at: (8 January 2006). 23 Stier, pp Wenjui Han and Jane Waldfogel, Child Care Costs and Women s Employment: A Comparison of Single and Married Mothers with Pre-School-Aged Children, SSQ, 82, 3 (2001), pp Michelle Budig, Are Women s Employment and Fertility Histories Interdependent? An Examination of Causal Order Using Event History Analysis, Social Science Research, 32, 3 (2003), pp Dwayne Benjamin and Michael Baker, The Role of the Family in Immigrants Labor Market Activity: An Evaluation of Alternative Explanations, American Economic Review, 87, 4 (1997), pp Jacob Alex Klerman and Arleen Leibowitz, Job Continuity among New Mothers, Demography, 36 (1999), pp Coleman, pp ibid 30 Robert D. Putnam, Making Democracy Work: Civic Traditions in Modern Italy, Princeton University Press, New Jersey, 1993, p Robert D. Putnam, Bowling Alone: America s Declining Social Capital, Journal of Democracy, 6, 1 (1995), pp

17 32 Mike Moffatt, What is the Labor Force Participation Rate? online, nd, available at: (11 April 2006). 33 U.S. Census PUMS. 34 Alan Agresti and Barbara Finlay, Statistical Methods for the Social Sciences, Prentice Hall, New Jersey, 1997, p David Garson, Logistic Regression, online, nd, available at: (11 May 2006) Sol Encel and Helen Studencki, Older workers: can they succeed in the job market? Australasian Journal of Ageing, 23, 1 (2004), pp Trond Petersen, Soc 271B Part II on Linear Regression Analysis, online, nd, available at: (15 April 2006). 39 U.S. Census PUMS U.S. Census Housing and Household Economic Statistics Division, online, nd, available at: (11 April 2006). 43 Grace X. Ma, Steven E. Shive, and Rosemary M. Feeley, The impact of acculturation on smoking in Asian American Homes, Journal of health care for the poor and underserved. 15 (2004), pp Garson, pp Thomas M. Carsey, POS 5737: Advanced Quantitative Methods, online, Fall 2003, available at: (10 May 2006). Urano/Social Capital and Immigrant Women s Employment 81