Printer Friendly

Intermarriage and the labor market outcomes of Asian women.

I. INTRODUCTION

Since the Immigration and Nationality Act of 1965 relaxed existing national origin quotas, the foreign-boRN population in the United States has surged. The growth in the immigrant population has sparked debates over the degree of economic assimilation of immigrants and factors affecting their assimilation (Borjas 1985, 1994; Lalonde and Topel 1992). More recently, the cultural assimilation of immigrants, as manifested by English proficiency, marriage between immigrants and natives, and child-rearing methods, has received academic attention.

There is no doubt that cultural and economic assimilations are linked. This article investigates one potential connection: the effect of intermarriage, defined as the marriage between a foreign-born individual and a native, on labor market outcomes of immigrant women, with a focus on Asians. Asians are a relatively new and fast-growing group in the United States and their population is expected to quadruple by 2050 (U.S. Census Bureau Population Projections 2010). Asians are considered successful in the U.S. labor market and in terms of human capital accumulation (Schoeni, McCarthy, and Vernez 1996). The role of the rise in marriages with natives in this success is an open question.

Intermarriage is often called the "final stage" in assimilation for ethnic minorities (Gordon 1964). The literature on intermarriage focuses on labor market outcomes of male immigrants, and generally groups all ethnicities together. If returns from inter- and intra-marriages differ by gender and ethnicity, existing evidence may not apply to all groups. Using data from the 2000 U.S. Census, this article discusses the labor market returns from intermarriage for Asian immigrant women, and compares them to non-Asian women.

Intermarriage can boost the earnings potential of the immigrant: this is the productivity hypothesis (Benham 1974). Marriage to a native can both compensate for an immigrant's lower income and encourage investment in host-country-specific capital, like learning a new language. Also, immigrants gain access to native networks which may be superior to immigrant networks for labor market prospects (Furtado and Theodoropoulos 2010).

Causal estimates of the effect of intermarriage on labor market outcomes are difficult to construct. An immigrant who is attractive to a native spouse can be positively selected into the labor market (Meng and Gregory 2005; Kantarevic 2004). Causality is further questioned if high earnings themselves determine intermarriage. Economists devised different methods to deal with selection and endogeneity, and found that intermarriage income premiums increased for men and children of immigrants in Australia and Europe (Meng and Gregory 2005; Meng and Meurs 2006; Qelikaksoy 2007; Gevrek 2009). This suggests that selection into intermarriage is negatively correlated with income. However, in the United States, upon controlling for selection, the intermarriage wage premium fell and became insignificant (Kantarevic 2004).

Intermarried Asian women and their spouses may differ from non-Asians. Despite being a high-skill group, cultural traditions such as patriarchy may discourage post-marital human capital investment by women. To control for biases, the article uses two instruments--the probability of intermarriage and the sex ratio, both calculated for a woman's age-group, metropolitan area, and birthplace. Even after controlling for unobservable heterogeneity, household structures between non-Asian and Asian women can differ post-marriage. Gender-based division of labor (or the specialization in home production by women) can imply different returns both from marriage and intermarriage for men and women. The degree can vary by ethnicity. Spousal income effect and investment in family stabilization can lead to different outcomes across intermarried and intra-married women.

In recent years, intermarriage rates for women in major immigrant groups have fallen (see figure 1). The rise in the stock of immigrants in the United States over the last decades allowed a larger share of immigrants to marry within their ethnic group (Chiswick and Houseworth 2011). On the other hand, ethnic exogamy between white or black natives and second- or third-generation Asian Americans and Mexican Americans is rising. Duncan and Trejo (2011) emphasize the role of intermarriages in ethnic attrition. Children born in mixed unions exhibit better socioeconomic outcomes than children of intra-marriages but are also less likely to identify with their immigrant parent's birthplace. Children of mixed-Asian ethnicity have increased in the last decade (Pew Research Center 2013). The role of natives in the family formation of Asian women, and the impact on female labor market outcomes will likely have consequences for inter-generation assimilation.

The current study begins with linear regression analysis to show that intermarried Asian women draw a significant wage penalty (3%). The raw intermarriage premium for the non-Asian sample is 23% and can be explained by selection. Subsequent instrumental variables (IV) estimates of the wage penalty for Asians are larger (about 20%), and non-Asians also face a penalty although estimates are imprecise. The article investigates whether a spousal income effect is at work, and finds that intermarriage penalties rise with husband's education for Asian wives. Assimilation patterns of intermarried Asians indicate that they have lower initial wages, market hours, and employment compared to intra-married Asians, but exhibit faster rates of growth over their years of stay.

The article is organized as follows. Section II presents a theoretical framework. Section III provides the econometric specification. Section IV introduces the data. Section V presents estimates and Section VI addresses the channels. Section VII discusses robustness checks. Section VIII concludes.

II. THEORETICAL BACKGROUND

The existence of a marriage premium has been extensively documented in the United States. Married men earn more than single men, even after controlling for observable human capital variables. Estimates of the premium vary, with typical cross-sectional estimates being about 10%-30%. (1) On the other hand, the existence of a marriage premium for women is contested.

The hypotheses that explain a marriage premium can be extended to an intermarriage premium. The productivity hypothesis argues that married people can accumulate more labor market specific human capital than single people through greater division of household labor (Becker 1973). A different explanation for the marriage premium is the selection hypothesis (Ginther and Zavodny 2001). More productive individuals may self-select into marriage.

There are many productivity gains from intermarriage. First, marrying a native is likely to improve English language proficiency of the immigrant. English fluency imparts significant wage gains in the U.S. labor market (Bleakley and Chin 2004). Secondly, many skills acquired in the immigrant's home country may not be transferable to the host country (Duleep and Regets 1997). A native spouse can accelerate the acquisition of host-country skills. Also, a native spouse grants an immigrant access to native networks. Costs associated with job search fall, and the probability of finding high-wage jobs rises (Aguilera 2002).

Benefits may not accrue to men and women equally. In a male-breadwinner, female-homemaker model, the household specialization of labor can be more advantageous for the husband's labor market prospects (Becker 1973). The gains from specialization increase if the differences in skills and abilities between the couple is larger (Becker 1981). In the United States, positive assortative mating by education is more common among couples with different ethnic backgrounds compared to co-ethnics (Chiswick and Houseworth 2011; Furtado 2012). If the correlation between husbands' and wives' skills differs by type of marriage, gains for the immigrant wives may differ as well.

Family division of labor can differ across Asian and non-Asian marriages for various reasons. The patriarchal authority of Asian men can be maintained or weakened by the process of migration (Espiritu 1999). If the Asian husband is expected to be the primary breadwinner for the family, high-earning women and low-earning men will have trouble finding spouses (Retherford, Ogawa, and Matsukura 2001). Women who quit their jobs upon marriage or childbirth may lose much of their labor market human capital, or may be discouraged from making further investments. If native spouses are more supportive of their wives' labor market efforts, there are more gains from intermarriage for the wife. Alternatively, if natives marry immigrant women from traditional societies to reemphasize gender-based household structures and child-rearing, the gains from intermarriage for Asian women are limited.

Intermarried individuals are not a random subsample of the population. There are many qualities that are valued both in the marriage and labor markets, such as IQ and beauty (Hammermesh and Biddle 2001; Meng and Gregory 2005).

More education, a longer time in the United States, and good English encourage intermarriage (Chiswick and Houseworth 2011; Kulczycki and Lobo 2002), and are also rewarded in the labor market. Factors like host-country attachment and motivation are difficult to measure objectively. Finally, people who work in high-paying jobs or attend higher education institutions in the United States are likely to have a larger network of native friends, high wages, and a native spouse.

Selection into intermarriage can vary by ethnicity. Asian Americans value career success (The Rise of Asian Americans, Pew Research Center 2013). Given the high returns to education in the United States, the more motivated and career-oriented women may intermarry if intermarriage benefits assimilation. The least traditional women may choose to marry outside their community, and these unobserved traits can be positively correlated with labor market success. Women with more attachment to the U.S. labor market may select a native mate. (2,3)

A third view of the marriage premium is that of signaling--marriage is a signal of productivity to employers (Korenman and Neumark 1991). Intermarriage can signal to employers that an immigrant will stay in the United States.

The previous discussion suggests labor market returns from intermarriage are mostly positive. Marriage can reinforce the gender-based division of household labor. This might be more severe for traditional Asian women. Furthermore, if Asian intermarried women have a comparative advantage in home production, as a special case of the exchange theory (Grossbard-Shechtman 1993), they may give up their labor market aspirations in exchange for societal status, legal residence in the United States, and better living conditions. The income effect of a high-earning husband can reduce a wife's wages. Women may postpone their job search or opt for "softer" jobs. Additionally, cultural clashes and feelings of alienation can affect the health and human capital of an immigrant. If communities react differently to the out-marriages of women and men, the effects can differ by gender. There may also be more avenues for labor market effort coordination among intra-married couples. This is the family investment hypothesis (Eckstein and Weiss 2002). Baker and Benjamin (1997) show that, in Canada, intra-married immigrant women initially work in low-paying, high-hours jobs to finance their husband's human capital investment in a credit-constrained labor market. Eventually both husbands and wives move to better jobs. Intermarried wives do not have to perform this borrowing function, since credit constraints are less binding for natives. (4) The presence of a "family investment" motive can be captured by comparing wage and hours assimilation profiles of intermarried and intra-married wives.

III. ECONOMETRIC STRATEGY

The main equation that identifies the effects of intermarriage on earnings is a traditional human capital equation with additional explanatory variables:

(1) log [W.sub.i] = [alpha] + [beta][Inter.sub.i] + [gamma][Z.sub.i] + [[epsilon].sub.i]

where log [W.sub.i] is the logarithm of hourly wages. Hourly wages are calculated by dividing annual wage and business income by annual hours worked. [Inter.sub.i] is a binary variable equal to one if the woman's husband was born in the United States, and zero if he was bom abroad. The article seeks to estimate an intermarriage premium, not a marriage premium. Estimations are run on a sample of married women, using appropriate person-level Census weights. Returns are calculated separately for Asian and non-Asian women.

Observable human capital, assimilation, and demographic characteristics are in the vector [Z.sub.i]. These include a set of education controls, a third-order polynomial in potential experience, veteran status, current Census region of residence indicators, a binary indicator for living in a metropolitan area, and the number of children. For the non-Asian sample, race and ethnicity controls--white, black, and Hispanic--are included. It is unclear whether assimilation variables like English proficiency, citizenship status, and duration of stay should be included in [Z.sub.i], since these variables can both "cause" and "be affected by" intermarriage. Nevertheless, given the importance of these variables in determining an immigrant's labor market outcomes, they are included as controls. Region of origin fixed effects are also included. For example, immigrants from Western Europe might find it easier to assimilate and share characteristics with the native population, improving their intermarriage rates.

The controls mentioned above account for observable heterogeneity between intermarried and intra-married women that can also generate variations in labor market productivity. Characteristics of the spouse help to understand the mechanisms that affect the wife's returns from intermarriage. Asian men typically have higher average earnings than native males, and intermarriage can reduce labor market returns for the Asian wife. The inclusion of spousal controls should eliminate a potential wage penalty from intermarriage for Asians. To test the direction of husband's influence, some specifications add controls for spousal wage, education, experience, and veteran status. If a spouse does not work, the wage is recoded as one cent, and a binary indicator of "nonworking" spouse is included.

An ordinary least squares (OLS) estimation of Equation (1) treats the marriage decision as exogenous. As discussed before, this may not be the case. A method commonly employed in the literature to correct for endogeneity and selection bias is the use of IV. (5) This article uses two instruments. The first instrument is the "probability of intermarriage" which shows the availability of mates from one's home country versus the availability of native partners in a woman's age-group and metropolitan area. (6)

(2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

[P.sub.i,c,m,a] is the probability of intermarriage for a woman i born in source country c, in age group a and residing in metropolitan area m. [UM.sub.c,m,a and [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] are the number of unmarried men from country c and the United States, respectively in age group a and metropolitan area m. Literature suggests that individuals are attracted to people of their own age, ethnic, and religious groups (Qian and Lichter 2001). People who belong to a larger group also tend to identify strongly with their ethnic group, increasing the chances of intra-marriage (Kalmijn and van Tubergen 2010). The geographic distribution of Asian and nonAsian immigrants across the United States is skewed. To reduce this skewness, the logarithm of the instrument is used.

Another factor affecting intermarriage for immigrant women could be competition from other women for eligible men of the same source country. An unbalanced sex ratio also decreases the overall chances of matrimony. A higher country-age-metropolitan area-specific female-to-male ratio facing woman i ([SR.sub.i,c,m,a]) lowers her chances of intra-marriage.

(3) [SR.sub.i,c,m,a] = ([F.sub.c,m,a] -1)/[M.sub.c,m.a]

where [F.sub.c,m,a] - 1 is the number of other women from country c in metro m and age group a (minus the immigrant woman herself). [M.sub.c,m.a] is similarly defined for men. Both the instruments are constructed using appropriate Census population weights.

The validity of the instruments relies on the assumption that "probability of intermarriage" and "sex ratios" affect a woman's wages only through the marriage decision. The "probability of intermarriage" instrument is closely related to the overall, as well as country-of-origin-specific, concentration of immigrants in the metropolitan area. In general, immigrants tend to congregate in large cities. If local economies are booming, wages are higher on average. Conversely, a recession or overcrowding in large cities can depress wages. As a check for economic conditions of the metropolitan area, the native unemployment rate corresponding to a person's age-group and metropolitan area is included in the main Equation (1). An indicator variable for living in a city is also included. (7)

The instrument may still be correlated with country-specific concentrations. The size of an immigrant group can affect the development of networks among its members (Edin, Fredriksson, and Aslund 2003). Living in an enclave can also impede assimilation. Enclaves have larger concentrations of undocumented immigrants and can increase a sense of insecurity among its occupants. If immigrants work in country-specific preferred occupational niches, wages of immigrants can be affected (Patel and Vella 2013). Immigrants who work in family occupations may not be in occupations of comparative advantage. Increased flow of entrant immigrants can exert a downward pressure on wages for existing immigrants. A control for the proportion of immigrants from one's country in the overall immigrant population of the metropolitan area is introduced. Variation in the "probability of intermarriage" instrument is now conditional on the relative group size in the metropolitan area. (8)

Industry composition variables can respond to sex ratios in a metropolitan area. Average wages in a metropolitan area can be higher although female wages fall, if the area has more male-dominated jobs which are typically higher-paying. Another concern with these instruments is that high-earning women may move to areas in response to differential probabilities of intermarriage or sex ratios. Internal migration research shows people move mostly as a response to economic opportunities. The average native female-to-male wage ratio for a person's age-group and metropolitan area is included to mitigate these concerns.

The metropolitan area controls described are also included in the main specification (1). The instruments are constructed only for identifiable metropolitan areas and IV estimates are presented for this sample. Linear regression results are reported for the entire sample. (9)

IV. DATA

A. Sample Selection

The analysis in the article uses data from the 5% sample of the 2000 U.S. Census, specifically the Integrated Public Use Microsample Series (Ruggles et al. 2013). The U.S. Census is particularly suited to the study of immigrant outcomes due to its large sample size and rich measures of information on race, ethnicity, and source countries of immigrants. The Census provides data on earnings, hours and weeks worked, as well as human capital variables like education, age, and English proficiency.

The Census is a cross-sectional dataset and a person is observed once. The Census allows creation of marital households, if husband and wife live together. An intermarriage is defined as a legal union between an immigrant woman and a native man. (10) The sample includes married immigrant women between the ages of 25 and 65. In order to study the effects of intermarriage on civilian labor market outcomes, I further restrict the sample to civilian female labor force participants who report earning a strictly positive income last year. (11) In 2000, labor force participation of foreign-bom women is below that of native women. (12) While labor force participation decisions and the differences by type of marriage, are not modeled, trends and implications are discussed later.

The 2000 Census only reports the incidence of marriage and not the year or number of marriages. Ideally, to identify the effect of intermarriage, the woman should marry after arriving in the United States. Structural factors influencing marriage and labor market outcomes differ across countries.

The 2000 Census reports an immigrant's year of arrival and current age. The sample is restricted to "young entrants" who immigrated at ages 23 or below. The assumption is that younger entrants arrive before their "marriage age." The median age of first marriage for native women in 2000 was 25.2 years. However, immigrants may marry at a younger age than natives. Simmons and Dye (2004) estimate the median age of marriage for all immigrant women to be 23.7 years in 2000, and for Asian women to be higher: 25.7 years. (13) The older age is explained by their propensity to attain tertiary education.

B. Descriptive Statistics

Table 1 presents the descriptive statistics of intra-married and intermarried immigrant women, as well as their husbands. The rate of intermarriage among Asian women (25%) is lower than non-Asians (about 29%). (14) Asian women, irrespective of type of marriage, have better human capital and labor market outcomes than non-Asians. Women who intermarry are more educated, older, and speak better English. In line with these characteristics, non-Asian intermarried women have higher wages, are more likely to be employed, and work in the high-skill sector. (15) Selection bias is, therefore, a concern. However, comparing among Asians, intermarried women have lower wages, earning 70 cents less per hour on average. The probability of employment or high-skill employment does not differ significantly by type of marriage. Spouse quality is not worse in an Asian intermarriage. Native husbands are better educated and older than immigrant husbands of Asian women and have comparable labor market outcomes. Intermarriages are more likely in military families--almost 40% of native husbands have some service experience. Finally, the correlation between spousal human capital characteristics is lower for intermarriages, and lowest for Asian intermarriages.

The absence of labor market benefits from intermarriage is not an Asian outcome. Asian immigrant men married to native women exhibit better human capital and labor market outcomes

(Table SI, Supporting Information). Their native wives also exhibit better human capital and market outcomes than the immigrant wives, although wives earn less than husbands. Correlation in spousal skills is still the lowest for intermarriages; however, the immigrant husband can now be expected to benefit from gender specialization.

V. ESTIMATION RESULTS

A. OLS Estimates of the Intermarriage Returns on Hourly Wages

Results from linear regression analysis of Equation (1), for both the Asian and non-Asian female subsamples, are shown in Table 2. The raw intermarriage return is large and positive for non-Asian women (+23%), and insignificant and small for Asian women (0.35%). Once human capital, demographic, and assimilation controls are added, Asian women face a significant 3.4% wage penalty. The raw return for non-Asian women falls sharply, and is now insignificant (about 0.6%). This confirms the hypothesis that intermarried women have observable characteristics that are valued in the U.S. labor market. (16)

Next, characteristics of the husband are added to the equation. If natives on average earn less than Asians, controlling for husband's characteristics should reduce the Asian intermarriage wage penalty. When spousal education, wages, and experience are added to the equation, the intermarriage wage penalty rises to 3.7% for Asian women (not shown in table). One might be tempted to consider this as "better spousal quality" among intermarriages but selection on unobservables and reverse causality have not been addressed. Adding a control for husband's veteran status along with other spousal controls reduces the intermarriage penalty to 2% for Asian women. Asian and non-Asian wives of ex-military personnel earn significantly less than civilian wives. Finally, the addition of metropolitan area controls to the models does not change coefficients significantly. For all models, the intermarriage coefficients for the non-Asian and Asian samples are significantly different.

Table S2 compares the Asian and non-Asian married male samples. Addition of controls render differences across the samples insignificant. Hence, the rest of the article focuses on married women. (17)

As mentioned before, the wives sample includes "young" entrants who arrived in the United States before age 24. Further dividing the sample into younger and older entrants, the wage penalty is insignificant for Asian women who arrived at ages below 18, and significantly negative for those who entered at ages 18-23. (18) A concern with younger entrants is that they have higher lifetime incomes via easier language acquisition (Bleakley and Chin 2004) and ability to assimilate in the host country. Age-of-entry is not exogenously determined for childhood entrants, it is correlated with family fixed effects (Bohlmark 2008). The benefits of intermarriage may be limited for young entrants. Alternatively, raising the age-of-entry can imply that results are skewed by people who marry abroad. I exclude all couples who arrived in the same year. Wage results are robust. Another concern is that immigrants may marry American military personnel stationed overseas and then move to the United States. Table 1 shows a proclivity for intermarriages in military households.

The exclusion of military couples reduces the intermarriage penalty for the Asian sample and increases the premium for non-Asians, but the difference between the groups is still significant at the 10% level.

B. IV Analysis

IV Estimates of the Intermarriage Returns on Hourly Wages. The wage penalty of Asian wives can be a result of negative selection on unobservable traits. If less-career-oriented Asians decide to intermarry, their labor market returns will be lower. Alternatively, the earnings distribution within the family could be decided before the marriage decision was undertaken. Using the instruments defined in Equations (2) and (3), I employ a 2SLS method and endogenize the marriage decision.

Asians often form smaller proportions of the total population of a metropolitan area; they are also more likely to live on the East or West Coast. To reduce the sensitivity of the instruments to small sample variations, only identifiable metropolitan areas with at least 500 people from one's country-of-origin are included. The first and second stage estimates from models with a full set of own, spousal, and metropolitan area controls are shown in Table 3.

The instruments have the expected signs. As the number of own-country women per men increases by one, the chances of intermarriage rise by 3.7%-4.9%. A greater availability of unmarried men from one's own country, relative to native men, reduces the probability of intermarriage. A 10% increase in the availability of unmarried men in a non-Asian woman's age-metropolitan area group reduces her probability of intermarriage by 3.3%, and the decrease is 2.7% for Asians. Coefficients are significant at the 1% level.

The model passes the under-identification and weak instruments tests. The values of the Kleibergen-Paap [chi square] test and the F-test of excluded instruments are sufficiently high to reject the respective nulls of an under-identified model and an identified model that suffers from a weak correlation between the instruments and the endogenous variable.

Second-stage IV estimates are presented in columns 2 and 4. Standard errors, clustered by age, metropolitan area, and country of birth cells are shown in parentheses. The wage penalty for Asian women increases to -23.9%. IV estimates are more negative than OLS estimates, implying that conditional on unobservables, selection into the labor market for intermarried Asians is positive. Taking unobservable selection into account also creates a wage penalty for non-Asians (-14.6%). The point estimate is smaller than that of non-Asians, but the difference between the groups is no longer significant, due to the large standard errors. (19) Further investigation of the spousal variables shows that education of the husband has a significant and positive impact on the probability of intermarriage and the wages of his wife--for the non-Asian sample, but not for Asian women.

Discussion of IV Estimates. Despite the inclusion of the metropolitan-area controls discussed at the end of Section III, the second-stage Hansen J-statistic (Table 3) shows that validity of instruments remains a concern for the Asian subsample. This section addresses further concerns about the instrument. All additional IV estimates discussed are available upon request.

The use of instruments from the 2000 Census implicitly assumes that structural factors affecting marriage conditions remained the same between 2000 and the actual year of marriage. (20) The sample is restricted to ages 25-40. The assumption is that the year 2000 most closely resembles marriage market conditions of this younger age-group. The gap between the penalties of Asian and non-Asians narrows, though the Asian wage penalty is larger. (21)

Intra-marriage might be with immigrants who are not from one's country of birth. The probability of intermarriage instrument is modified to include all immigrants in one's age group-metropolitan area cell. (22) Results are robust to the use of this alternate instrument.

Finally, some immigrants may have married prior to migration. The sample is restricted to those entering at ages below 18. Intermarriage penalties are now similar across the two groups although the Asian penalty is still larger. Conditional on observables, selection into intermarriage for Asian and non-Asian younger entrants could be more similar.

One explanation for the larger IV estimates for both groups may be that the instruments are unable to generate enough variation to identify the marriage equation from the wage equation. Alternatively, there is a possible selection into work bias. Wages are only available for workers. This article does not explicitly address the labor force participation decision for different groups of women. Participation rates vary by type of marriage, and within type of marriage by level of education. (23) Results are not shown, but the linear intermarriage impact on labor force participation, conditional on all observables, is 2.8% for non-Asian women and 0.9% for Asians. However, the IV estimates show that intra-married non-Asians are more likely to enter the labor market. Modeling participation trends can further increase the difference between the Asian and non-Asian samples.

IV estimates are robust to the different specifications. The intermarriage penalty for the Asian sample continues to be larger than the nonAsian sample. However, there still may be concerns about the instruments. Also, OLS estimates are more conservative about the intermarriage penalty. I will focus on linear estimates for the rest of the article.

VI. EXPLANATIONS FOR THE ASIAN INTERMARRIAGE WAGE PENALTY

A. Intermarriage Returns on Labor Market Employment

Selection into employment can potentially explain the negative wage penalty. If intramarried Asians have traditional husbands, these women may seek employment only when they command a high wage. Intermarried women may be less selective and accept low-wage employment. If this were true, I expect an intermarriage premium for employment but not wages for Asian women. Table 4 shows the effects of intermarriage on employment for Asian and non-Asian women. I see no intermarriage premium on employment for Asians. Conditional on own and spousal observable characteristics, non-Asian women enjoy about a 1% intermarriage employment premium. Higher employment from intermarriage is thought to be related to access to native networks (Furtado and Theodoropoulos 2010).

The Asian intermarriage wage penalty can result from working in soft or part-time jobs. If Asians have high-earning husbands, it is not implausible to assume that the wives are working fewer hours. Table 4 also shows the effects of intermarriage on labor market hours. Labor force participants earning strictly positive wages and working non-zero market hours are included. Intermarried women, indeed, work fewer hours. While the raw penalty on hours worked is higher for non-Asians, there is no statistical difference in the intermarriage penalty between the two samples, once observable controls are added.

A male-breadwinner model, motivated by a larger endowment of human capital to the husband, would imply the more educated spouse earns higher wages and works more. Women married to natives with higher earnings potential, conditional on their own observable traits, have a smaller intermarriage premium. This is explored in the next section.

B. Heterogeneous Returns from Intermarriage

A possible explanation for the wage penalty faced by Asian intermarried women is a spousal income effect. Are labor market efforts or investments in human capital lower as a result of the high income potential of their native spouses? This section re-estimates Equation (1) by stratifying the sample on the basis of husband's education. Education is a well-known indicator of a person's human capital and labor market productivity. The correlation between own and spousal education is low in Asian intermarriages (Table 1). The sample is separated by husband's skill level. The idea is to capture heterogeneous returns (3t. from intermarriage by husband type, conditional on the wife's education and other own observable traits. Five spousal education categories e are used: high school dropout, high school graduate, some college, associate degree or Bachelor's degree, and graduate or professional education. (24)

OLS estimates of the heterogeneous returns from intermarriage, by type of husband, are shown in Table 5. The intermarriage wage premium is highest for Asian wives of men with low levels of education--high-school dropouts. The penalty appears for women married to men with tertiary education. The wage penalty for Asian women married to men with at least some college education persists even after wife's characteristics are included in the specification. The biggest disadvantages for the non-Asian sample appear for women married to husbands with less than college graduate degrees. Finally, husband's wage and experience are added to the model. The wage penalty is reduced only for Asians married to men with higher levels of education, further providing evidence of the income effect.

Intermarriage has a negative effect on labor market hours of the wife, across all education levels of the husband (Table 6). For all models and both samples, I cannot reject the hypothesis that the penalty on hours is the same across all levels of husband's education. IV estimates are provided in the Supporting Information.

At lower levels of education, both husband and wife may need to contribute to family finances. Intermarriage with a native, who has access to native networks, can substitute for lack of skill and generate a premium at this end of the distribution. Asian women are disproportionately represented in the upper end of the skill distribution and the negative income effect of having a high-earning husband is dominant.

C. Wage and Hours Assimilation in an Intermarriage

Intra-married households can have more avenues for coordinated labor market efforts. Owing to credit constraints, immigrant women may take on worse jobs, with low pay and longer hours, to assist the human capital formation of their immigrant husbands. This occurs at the time of entry into the host country. Once husbands have built their capital, the wives move to better jobs (Baker and Benjamin 1997). Since natives are less credit-constrained, their wives do not have to perform a borrowing function for the family. The family investment behavior can result in an intra-marriage premium. Asian families are less likely to be credit-constrained due to their higher earnings and the family investment on part of wives could be weaker.

Alternatively, if native men choose Asian wives for a supposed comparative advantage in family-building, these wives invest more in family and less in labor market work at the beginning of their marriage. Over time, they move to better and higher-paying jobs.

To analyze initial differences in labor supply and wages, and the growth in the variables over time and by type of marriage, I estimate Equation (4) on a pooled sample of married female immigrants from the 2000 Census and the 2007 American Community Survey. (25)

(4) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

[D.sub.i,t] is the dependent labor market variable. (26,27) [Cohort.sub.i,c] are dummies for six immigrant cohorts. Cohorts are created on the basis of an immigrant's year of arrival to the United States--1940-1949, 1950-1959, 1960-1969, 1970-1979, 1980-1989, and 1990-2000. The newest cohort is the base group. [[beta].sub.c] captures the cohort effect for cohort c. [[beta].sup.Inter] shows the difference in initial conditions between intramarried and intermarried women of cohort c. [WUSA.sub.i,t] measures years in the United States, [gamma] is an assimilation parameter for all immigrant women. [[gamma].sub.Inter] measures the difference in growth by type of marriage. (28) [Z.sub.i,t] is a vector of human capital and demographic characteristics similar to Equation (1). [[delta].sub.t] is a time dummy common for intermarried and intra-married women. This is a necessary assumption to identify both assimilation and cohort effects. Husband's years of education and a quadratic expression of his U.S. labor market experience are also included in [Z.sub.i,t]. (29)

Following Baker and Benjamin (1997), the assimilation patterns of a representative cohort--those who arrived between 1980 and 1989--are mapped, separately for Asians and non-Asians (Figure 2). (30) All variables, except cohort effects and years in United States, are set to zero. Our estimates do not indicate the presence of a "family investment" motive in intra-married households; this has also been reported by Blau et al. (2003) in the context of the United States. As shown in graph (a), initial wages of intermarried Asians, not intra-married Asians, are lower. Intermarried Asians exhibit faster wage assimilation. Employment growth is also faster for intermarried Asians (graph c). Next, graphs (b) and (d) show the wage and employment assimilation for non-Asians: intermarried women always earned more and had higher employment probabilities. There is positive assimilation in hours worked for all groups of women, though intermarried women always work less hours (graphs e and f).

The lower levels of initial labor market involvement and subsequent growth for intermarried Asians could result from an extended job search. The income buffer of having a high-earning native husband can allow a wife to undertake longer searches. It is not clear why, conditional on their observable characteristics, intermarried Asians need more time to search, compared to non-Asians.

On the other hand, if Asian women initially invested more in family and eventually switched to high-wage employment, similar assimilation patterns emerge. (31) Without time-use data, it is difficult to identify the exact source of the assimilation patterns. As an indirect test, I considered 25-45-year-old women with only one nonadult child. Intermarriage employment penalties were highest for Asian mothers of infants and children aged 5-9. Wage penalties from intermarriage were U-shaped, with mothers of 5-9-year-old children and pre-teens facing a 8% penalty. Mothers of infants or teenagers had insignificant penalties. Asian women may quit work following childbirth, and take lower-paying jobs when their children are in school. Non-Asian mothers of teenagers experienced the largest wage loss.

VII. ROBUSTNESS CHECKS

A. Heterogeneous Asian Subethnicities

The term "Asian" encompasses a number of subethnicities. If the intermarriage premium differs between Asians and non-Asians, it is not implausible that returns are heterogeneous across Asian subgroups. Four Asian subcategories are considered--Southeast Asia, East Asia, Indian Subcontinent, and Islamic Nations of the Middle East. The rates of intermarriage are different.

About 30%-34% of East and Southeast Asian women intermarry. Only 6% of women from the Indian subcontinent and 16% of women from the Islamic nations intermarry.

For each of these subgroups, Equation (1) is estimated and presented in Table S5. The raw intermarriage premium would imply that all the penalties originate from the East Asian group. Intermarried women from the Indian subcontinent and the Middle East make a 13% premium. Once the wife's observable characteristics are added to the estimation, the premium for these two groups is no longer visible. A joint test of significance of the coefficients rejects that the coefficients are different.

Nevertheless, the point estimates imply that penalties are smaller for the groups that are "less permissive" about intermarriage. If the decision to intermarry is taken by the less traditional women, they are unlikely to give up their labor market options and opt for a gender-based home production role. It is also possible that natives hold the promise of labor market employment which is not possible in a traditional immigrant family set-up. It has been discussed that the Asian intermarriage wage penalty could result from the fact that skilled immigrant women married to more traditional immigrant husbands were selecting into employment only when they commanded a high wage. Even for the so-called less permissive groups, the intermarriage returns to employment are negligible.

B. Heterogeneous Race Type of the Husband

This article defines intermarriage as marriage between an immigrant and a native. In the context of an Asian intermarriage, the union of an Asian woman and a second-generation Asian man born in the United States, is an intermarriage. The union of an Asian woman and a native Caucasian or African American man is also an intermarriage. It is possible that the returns to marriages between co-ethnics, even if they are born in different countries, might be different from marriages between people of different ethnicities. The negative return could be a result of cultural alienation that has a detrimental impact on the human capital accumulation and labor market productivity of the immigrant wife.

Table S6 shows the returns from intermarriage on log hourly wages for immigrant wives, differentiated by husband's ethnicity. Native husbands of four mutually exclusive ethnic groups are considered--non-Hispanic white, non-Hispanic black, Asian, and Hispanic. The base group for comparison is intra-marriages between Asian women and Asian immigrant men, since intra-marriages with other immigrants is rare. Eighty percent of Asian intermarriages are with a Caucasian husband and 13% to native-born Asian men.

There is obvious heterogeneity in wives across the different husband ethnicities. There is a large raw wage penalty from marrying a native black man, and there is a high premium for marrying a co-ethnic. When wife and husband characteristics are controlled, the familiar result of a large intermarriage penalty across most husband ethnicities re-emerges. The penalties are not statistically different. When spousal controls are added, the fall in the penalty is largest for women married to white husbands. The income effect of marriage to "culturally dissimilar" husbands may be larger.

VIII. CONCLUSION

Intermarriage is considered to be a step forward in the assimilation process for ethnic minorities. The impact of cultural assimilation need not be homogeneous across sexes or ethnic groups of immigrants. Using both least squares and IV methods, this article shows that Asian women married to natives face a significant wage penalty. This is not true of intermarried non-Asian women, or even Asian men. I examine the mechanisms via which educated women married to skilled men could receive penalties. Wage penalties of Asian women, but not non-Asians, increase with the level of husband's education. There might be a spousal income effect that discourages labor market efforts of intermarried Asian women.

Family investment on the part of intra-married Asian women does not explain the intermarriage penalty. On the contrary, I find intermarried women exhibiting low initial wage and employment levels, and subsequently a faster growth in these variables compared to intra-married Asians. Either intermarried Asians substitute time and effort away from the labor market to stabilize the family and eventually move to better jobs, or their job search period takes longer than non-Asian intermarried women.

It would be of further interest to explore the differences in the kind of jobs intermarried Asian and non-Asian women tend to hold. Longitudinal data would provide an opportunity to better deal with selection issues, as well as allow us to follow the work history of intermarried women.

Time-use data are also excellent tools to analyze the efforts of Asian women in the early phases of their marriages.

The success of Asians, compared to other immigrant groups, is a story of averages. For Asian women who intermarry, cultural assimilation appears to occur without economic assimilation. The results in this article have implications for the inter-generational assimilation of Asian children. There is a wealth of evidence supporting that mother's education boosts educational attainment of the offspring (Behrman 1997). On the other hand, if the Asian intermarriage penalty confirms gender roles in the household, it is also likely to affect marriage satisfaction and probability of divorce (Bertrand, Kamenica, and Pan 2015). The effects of having a well-educated mother, whose labor market outcomes are not on par with her skills, on the socioeconomic outcomes of the next generation can be very complex and are worth investigating.

ABBREVIATIONS

IV: Instrumental Variables

OLS: Ordinary Least Squares

doi: 10.1111/ecin.12229

REFERENCES

Aguilera, M. "The Impact of Social Capital on Labor Force Participation: Evidence from the 2000 Social Capital Benchmark Survey." Social Science Quarterly, 83(3), 2002, 853-74.

Baker, M., and D. Benjamin. "The Role of the Family in Immigrants' Labor Market Activity: An Evaluation of Alternative Explanations." American Economic Review, 8(4), 1997, 705-27.

Becker, G. S. "A Theory of Marriage: Part I." Journal of Political Economy, 81(4), 1973, 813-46.

--. A Treatise on the Family. Cambridge, MA: Harvard University Press, 1981.

Behrman, J. "Mother's Schooling and Child Education: A Survey." Penn Institute for Economic Research Working Paper 97-025, Department of Economics, University of Pennsylvania, 1997.

Benham, L. "Benefits of Women's Education Within Marriage." Journal of Political Economy, 82(2), 1974, S57-71.

Bertrand, M., E. Kamenica, and J. Pan. "Gender Identity and Relative Income Within Households." Quarterly Journal of Economics, 130(2), 2015, 571-614.

Blau, F. D., L. M. Kahn, J. Y. Moriarty, and A. Souza. "The Role of the Family in Immigrants' Labor-Market Activity: An Evaluation of Alternative Explanations: Comment." American Economic Review, 93(1), 2003, 429-47.

Bleakley, H., and A. Chin. "Language Skills and Earnings: Evidence from Childhood Immigrants." Review of Economics and Statistics, 86(2), 2004, 481-96.

Bohlmark, A. "Age at Immigration and School Performance: A Siblings Analysis Using Swedish Register Data." Labour Economics, 15(6), 2008, 1366-87.

Borjas. G. "Assimilation, Changes in Cohort Quality, and the Earnings of Immigrants." Journal of Labor Economics, 3(4), 1985,463-89.

--. "The Economics of Immigration." Journal of Economic Literature, 32(4), 1994, 1667-717.

Bratsberg, B., J. Ragan, Jr., and Z. Nasir. "The Effect of Naturalization on Wage Growth: A Panel Study of Young Male Immigrants." Journal of Labor Economics, 20(3), 2002, 568-97.

Card, D. "Immigration and Inequality." American Economic Review: Papers and Proceedings, 99(2), 2009, 1-21.

Celikaksoy, A. "A Wage Premium or Penalty: An Analysis of Endogamous Marriage Effects among the Children of Immigrants?" Danish Journal of Economics, 145(3), 2007, 288-311.

Chiswick, B., and C. Houseworth. "Ethnic Intermarriage among Immigrants: Human Capital and Assortative Mating." Review of Economics of the Household, 9(2), 2011, 149-80.

Duleep, H., and M. Regets. "The Decline in Immigrant Entry Earnings: Less Transferable Skills or Lower Ability?" Quarterly Review of Economics and Finance, 37(S), 1997,189-208.

Duncan, B., and S. Trejo. "Intermarriage and the Intergenerational Transmission of Ethnic Identity and Human Capital for Mexican Americans." Journal of Labor Economics, 29(2), 2011, 195-227.

Eckstein, Z., and Y. Weiss. "The Integration of Immigrants from the Former Soviet Union in the Israeli Labor Market," in The Israeli Economy, 1985-1998: From Government Intervention to Market Economics, Essays in Memory of Prof. Michael Bruno. Cambridge, MA: MIT Press, 2002, 349-78.

Edin, P., P. Fredriksson, and O. Aslund. "Ethnic Enclaves and the Economic Success of Immigrants--Evidence from a Natural Experiment." Quarterly Journal of Economics, 118(1), 2003, 329-57.

Espiritu, Y. "Gender and Labor in Asian Immigrant Families." American Behavioral Scientist, 42(4), 1999, 628-47.

Furtado. D. "Human Capital and Interethnic Marriage Decisions." Economic Inquiry, 50(1), 2012, 82-93.

Furtado, D., and N. Theodoropoulos. "Why Does Intermarriage Increase Immigrant Employment? The Role of Networks." The B.E. Journal of Economic Analysis and Policy; Topics, 10(1), 2010, 1-33.

Gevrek, Z. "Interethnic Marriage and the Labor Market Integration of Immigrants." mimeo. University of Konstanz, 2009.

Ginther, D., and M. Zavodny. "Is the Male Marriage Premium Due to Selection? The Effect of Shotgun Weddings on the Return to Marriage." Journal of Population Economics, 14(2), 2001,313-28.

Gordon, M. M. Assimilation in American Life: The Role of Race, Religion and National Origins. New York: Oxford University Press, 1964.

Grossbard-Shechtman, S. On the Economics of Marriage: A Theory of Marriage Labor, and Divorce. Boulder, CO: Westview Press, 1993.

Hammermesh, D., and J. Biddle. "Beauty and the Labor Market." American Economic Review, 84(5), 2001, 1174-94.

Kalmijn, M., and F. van Tubergen. "A Comparative Perspective on Intermarriage: Explaining Differences in Marriage Choices among National Origin Groups in the United States." Demography, 47(2), 2010, 459-79.

Kantarevic, J. "Interethnic Marriages and Economic Assimilation of Immigrants." IZA-Institute for the Study of Labor, Discussion Paper Series No. 1142, Bonn, May, 2004.

Korenman, S., and D. Neumark. "Does Marriage Really Make Men More Productive?" Journal of Human Resources, 26(2), 1991,282-307.

Kulczycki, A., and A. Lobo. "Patterns, Determinants and Implications of Intermarriage among Arab Americans." Journal of Marriage and Family, 64(1), 2002, 202-10.

LaLonde, R. J., and R. H. Topel. "The Assimilation of Immigrants in the U. S. Labor Market. Immigration and the Workforce: Economic Consequences for the United States and Source Areas, " in Immigration and the Workforce: Economic Consequences for the United States and Source Areas, edited by G. Borjas and R. Freeman. Chicago: University of Chicago Press, 1992.

Loh, E. S. "Productivity Differences and the Marriage Wage Premium for White Males." Journal of Human Resources, 31(3), 1996, 566-89.

Meng, X., and R. G. Gregory. "Intermarriage and the Economic Assimilation of Immigrants." Journal of Labor Economics, 23(1), 2005, 135-76.

Meng, X., and D. Meurs. "Intermarriage, Language, and Economic Assimilation Process: A Case Study of France." International Journal of Manpower, 30(2), 2006, 127-44.

Nottmeyer, O. "Does Intermarriage Pay Off? A Panel Data Analysis." IZA-Institute for the Study of Labor, Discussion Paper Series No. 5104, Bonn, May 2010.

Patel, K., and F. Vella. "Immigrant Networks and Their Implications for Occupational Choice and Wages." Review of Economics and Statistics, 95(4), 2013, 1249-77.

Pew Research Center. "The Rise of Intermarriage: Rates, Characteristics Vary by Race and Gender." 2012. http:///www.pewsocialtrends.org/files/2012/02/SDTIntermarriage-II.pdf

--. "The Rise of Asian Americans." 2013. http://www.pewsocialtrends.org/files/2013/04/AsianAmericans-new-full-report-04-2013.pdf

Qian, Z., and D. Lichter. "Measuring Marital Assimilation: Intermarriage among Natives and Immigrants." Social Science Research, 30(2), 2001,289-312.

Retherford, R. D.. N. Ogawa, and R. Matsukura. "Late Marriage and Less Marriage in Japan." Population and Development Review, 27(1), 2001, 65-102.

Ruggles, S., J. T. Alexander, K. Genadek, R. Goeken, M. B. Schroeder, and M. Sobek. Integrated Public Use Microdata Series: Version 5.0 [Machine-readable database], Minneapolis: University of Minnesota, 2013.

Schoeni, R., K. McCarthy, and G. Vemez. The Mixed Economic Progress of Immigrants. Santa Monica, CA: Rand, 1996.

Simmons, T., and J. Dye. "What Has Happened to Median Age at First Marriage Data?" Paper presented at the Annual Meeting of the American Sociological Association, Population Division, U.S. Census Bureau, 2004. https://www.census.gov/acs/www/Downloads/library/ 2004/2004 Simmons 01.pdf

U.S. Census Bureau. "Population Projections for 2050." 2010. http://www.census.gov/newsroom/releases/arc hives/population/cb 12-243.html

SUPPORTING INFORMATION

Additional Supporting Information may be found in the online version of this article:

Table S1. Descriptive Statistics of Non-Asian and Asian Husbands and Their Spouses

Table S2. OLS Effects of Intermarriage on Log Hourly Wage of Immigrant Men

Table S3. IV Estimates of Intermarriage on Employment and Hours of Immigrant Women

Table S4. Estimates from Equation (4): Assimilation Profiles of Labor Market Variables

Table S5. Effects of Intermarriage, Differentiated by Asian Sub-ethnicity

Table S6. Effects of Intermarriage, by Spousal Ethnic Group

(1.) Korenman and Neumark (1991) and Loh (1996) have surveyed the literature on this topic.

(2.) The path to legal permanent residence and citizenship is easier with marriage to a native. Citizenship expands job opportunities (Bratsberg, Ragan, and Nasir 2002).

(3.) Asian women, whose skills have higher returns in the United States compared to their home countries, can be expected to exhibit higher attachment to the U.S. labor market.

(4.) Blau et al. (2003) do not find evidence for coordinated labor market effort among immigrant couples in the United States. Immigrant wives and husbands arriving in the United States work and earn less upon arrival, compared to natives. Over time both sexes exhibit positive wage and hours assimilation.

(5.) Researchers have also used Heckman two-step method (Meng and Gregory 2005) and panel-data techniques (Nottmeyer 2010).

(6.) People are stratified into 5-year age brackets from which they choose a mate.

(7.) The 2000 Census provides the city codes.

(8.) Metropolitan area-wide macroeconomic factors might still affect group size and the probability of intermarriage, and in turn affect wages. I also checked the group size from 2010. Immigrant networks are stable across metropolitan areas and over time (Card 2009). Results are largely unchanged. Group size variables from earlier Censuses are not appropriate since metropolitan area demarcations have changed.

(9.) Linear point estimates for the metropolitan areas-only sample are smaller though not statistically different from the entire sample.

(10.) Cohabiting couples or same-sex marriages are not included in the sample.

(11.) The top and bottom 1 % of earners are removed from the sample to reduce the influence of outliers.

(12.) The participation rate is 70% for native women, 54% for immigrant women, and 60% for Asian women. Asians are one-fourth of the total immigrant female population, but one-third of married working women.

(13.) The American Community Survey 2008 is the year closest to 2000 when age at first marriage is identifiable. The median age of marriage for Asian female immigrants was about 26, and higher than that of natives.

(14.) According to the Pew Research Center (2012), Asian Americans have one of the highest propensities to "marry out" of their ethnicity. This term Asian American includes both native and foreign-born Asians, and both sexes. Overall foreign-bom are less likely to marry-out, and males are less likely than women.

(15.) Managerial, professional, and technical jobs are defined as high-skill occupations.

(16.) Returns to human capital and demographic variables follow the expected direction and magnitude.

(17.) The raw intermarriage premium for Asian men (8.7%) is smaller. Once own controls, spousal, and metropolitan controls are added, the intermarriage premium between

the two samples is not significantly different. The results of the Asian male sample do indicate that the role of family in Asian male and female intermarriages are different.

(18.) For some ages of entry below 18, Asian women still face a large intermarriage penalty. Small sample sizes reduce the precision of these estimates.

(19.) If metropolitan area controls were removed from the 2SLS estimation, non-Asians once again enjoy a wage premium. Asian women continue to have a significant wage penalty.

(20.) As discussed before, the 2000 Census does not report exact year of marriage.

(21.) The actual size of the penalties is smaller than those reported in Table 3. Younger married couples may have less traditional households and the penalties of intermarriage can be lower.

(22.) Less than 10% of intra-marriages for the Asian sample occur with members of another country. Hence this is not our preferred instrument. This instrument also fails the Hansen J test for the Asian sample.

(23.) The labor force participation gap between intermarried and intra-married Asian women is smaller (68% and 63% respectively), and this is seen across all levels of education.

(24.) 80% of intermarried Asians and 60% of non-Asians have husbands with at least some college experience.

(25.) Borjas (1985) shows that more than one nationally representative cross-section is needed to study cohort and assimilation effects jointly if the characteristics of cohorts of are changing over time.

(26.) Like Baker and Benjamin (1997), market hours are defined annually instead of hourly. Results are qualitatively similar for both measures. Annual hours capture the effects of full-time and part-time work.

(27.) Sample selection is similar to Section IV.A.

(28.) In the estimation, residence is a quadratic term.

(29.) I allow the spousal effects to vary by type of marriage. For an immigrant husband, his U.S. labor market experience is his years of stay or potential experience, whichever is smaller.

(30.) Estimates that these graphs are based on are in Table S4.

(31.) Can a similar pattern emerge if intermarried Asians, with a comparative advantage in home production, exchanged their labor market efforts for higher societal status and better living conditions in the United States? An advantage in home production usually trades-off with an advantage in the labor market. Not shown here, intermarried Asians have lower high-skill occupation employment at the beginning of the marriage, and eventually overtake intra-married Asians. Intermarried non-Asians always have higher high-skill sector employment. It is possible that intermarried Asians developed skills and human capital within the marriage which allowed them to move to the high-skill sector, but this is hard to ascertain given the limitations of our data. The assimilation patterns are robust to the exclusion of women who are still in school.

SUKANYA BASU, I am grateful to the seminar and conference participants at the University of Rochester, United States Naval Academy, Vassar College, Western Economic Association International Conference, New York State Economic Association Annual Meeting, and Norface Migration Conference "Migration: Global Development, New Frontiers." I thank Ronni Pavan, Michael Insler, and Sarah Pearlman for their suggestions with the article, and Noah Kulick for his research assistance. I would also like to thank the editor and an anonymous referee for their comments. All errors and omissions are mine.

Basu: Assistant Professor. Department of Economics, Vassar College, Poughkeepsie, NY 12604. Phone 1 -845-437 7016, Fax 1-845-437-7576, E-mail subasu@vassar.edu

TABLE 1

Descriptive Statistics of Non-Asian and Asian Wives and
Their Husbands

                                  (A) Characteristics of Wives

                              All Female Immigrants Asian Immigrants

                            Immigrant    Native    Immigrant   Native
                             Husband     Husband    Husband    Husband

Percentage                    71.05      28.95      74.9       25.1
Years of education            10.12      13.1       13.62      13.99
                              (4.37)     (3.22)     (3.95)     (3.5)
Age                           38         42.84      36.75      39.28
                              (9.96)    (11.19)     (8.45)     (9.61)
Bad English                   35.06       4.83      14.88       1.87
                             (47.9)     (21.76)    (35.44)    (13.76)
Military serv.                 1.66       2.87       1.36       2.51
                             (12.76)    (16.7)     (11.57)    (15.64)
Employed (if in LF)           91.1       96.33      96.8       96.3
                             (28.75)    (19.03)    (18.91)    (17.52)
Hourly wage (if working)      12.79      15.97      18.3       17.69
                              (9.56)    (10.8)     (12.05)    (11.63)
% in high-skill               21.65      40.26      45.4       45.36
  occupations                (40.95)    (48.9)     (49.79)    (49.79)
Market hours worked/week      35.94      36.41      38.22      38.68
                             (13.15)    (12.9)     (12.9)     (12.55)
Years in United States        21.41      30.12      18.94      24.55
                             (11.32)    (12.68)     (8.92)    (10.11)
Cor(own educ., spouse          0.5789     0.5354     0.6335     0.4851
  educ.)
Cor(own wage, wage)            0.2614     0.2513     0.3293     0.2682
N                             89,421     38,047     27,054      9,326

                                 (B) Characteristics of Husbands

                              All Female Immigrants Asian Immigrants

                            Immigrant   Native    Immigrant   Native
                             Husband    Husband    Husband    Husband

Percentage                    71.05      28.95      74.9       25.1
Years of education            10.08      13.9       14.54      14.96
                              (4.7)      (3.22)     (4.31)     (2.9)
Age                           41.14      45.71      40.9       42.25
                             (11.13)    (12.53)     (9.44)    (11.04)
Bad English                   29          1.6       11.58       0.84
                             (45.66)    (12.5)     (31.9)      (9.11)
Military serv.                 6.93      39.5        7.47      43.9
                             (25.39)    (48.87)    (26.63)    (49.64)
Employed (if in LF)           95.03      97.26      97.06      98.02
                             (21.92)    (16.69)    (17.1)     (14.3)
Hourly wage (if working)      15.8       21.61      22.52      23.17
                             (10.75)    (13.12)    (14.67)    (13.34)
% in high-skill               15.48      39.35      48.35      52.07
  occupations                (35.8)     (48.7)     (50)       (50)
Market hours worked/week      42.95      44.63      44.47      45.56
                             (12.03)    (12.1)     (12.73)    (11.76)
Years in United States        21.26       --        19.32       --
                             (11.37)                (9.37)
Cor(own educ., spouse
  educ.)
Cor(own wage, wage)
N                            89,421     38,047     27,054      9,326

Notes: Standard deviations in parentheses. Statistics use
appropriate person-level weights from the 2000 U.S. Census.
Source: Census 2000 5% sample.

TABLE 2

OLS Effects of Intermarriage on Log Hourly Wage of Immigrant Women

                              Non-Asian Women          Asian Women

Raw Estimates (no controls)     0.2304 ***               0.0035
                                 (0.0052)                (0.01)
[R.sup.2]                         0.0306                  0.001
  [chi square] from                            406.16
  difference of                                 0.00
  coefficients test p value
Controls: Educ., Exp.,            0.0058               -0.0336 ***
  children, English, Vet.         (0.006)                (0.01)
  Status Citizen, Years in
  United States, Ethnicity,
  Census Reg., birthplace
[R.sup.2]                         0.2763                 0.2613
[chi square]  from                             11.32
  difference of                                0.0008
  coefficients test p value
Controls: Row 2 + Spousal         0.0042               -0.0199 **
  Wage, Vet. Status,             (0.0062)               (0.0104)
  Education, and Experience
[R.sup.2]                         0.2787                 0.2643
[chi square]  from                              3.9
  difference of                                0.0481
  coefficients test p value
Controls: Row2 + Row3 +           0.0043                -0.0195 *
  Metropolitan                   (0.0062)               (0.0105)
  Characteristics
[R.sup.2]                          0.282                 0.2661
[chi square]  from                              3.78
  difference of                                0.0517
  coefficients test p value
N                                 61,141                 21,595

Notes: The dependent variable is log hourly wage. The
explanatory variable of interest is a binary variable = 1 if
husband is U.S.-born and = 0 if husband is an immigrant.
Robust standard errors are reported in parentheses.
Estimations use appropriate person-level weights from the
2000 U.S. census.

* Significant at 10%; ** significant at 5%;
*** significant at 1%.

Source'. Census 2000 5% sample.

TABLE 3

IV Estimates of Intermarriage on Log Hourly Wage of Immigrant Women

                                 Non-Asian Women
                                   First Stage      Second Stage

Intermarriage                          --             -0.146 **
                                                       (0.072)
Ability to speak English well      0.1104 ***         0.111 ***
                                    (0.0051)          (0.0106)
Years in the United States         0.0094 ***         0.016 ***
                                     (0.001)          (0.0014)
Native unemployment                -0.0249 ***       -0.0674 ***
                                    (0.0083)           (0.012)
Female/Male wage ratio              0.206 ***        -0.194 ***
                                     (0.048)           (0.058)

Own group size in metro             0.159 ***        -0.126 ***
                                    (0.0197)           (0.174)
Husband's wage                     0.0001 ***        0.0002 ***
                                    (0.00002)         (0.00004)
Husband's experience               -0.0046 ***       -0.0044 ***
                                    (0.0004)          (0.0015)
Husband's years of education       0.0152 ***        0.0085 ***
                                    (0.0006)          (0.0015)
Husband's veteran status            0.295 ***          0.0276
                                    (0.0078)           (0.023)
Education dummies                      Yes               Yes
U.S. census region dummies             Yes               Yes
Probability of intermarriage       -0.033 ***            --
                                    (0.0028)             --
Sex ratio                          0.0367 ***            --
                                    (0.0033)             --
N                                    45,730            53,171
F-test value of excluded             120.62      Hansen's   J 0.13
instruments
Under-identification text            161.87       p-value   0.9084
  [chi square] value

                                   Asian Women
                                   First Stage       Second Stage

Intermarriage                          --             -0.239 ***
                                                        (0.09)
Ability to speak English well       0.127 ***         0.124 ***
                                    (0.0068)           (0.0154)
Years in the United States          0.009 ***         0.0098 ***
                                     (0.002)           (0.0024)
Native unemployment                  -0.004           -0.111 ***
                                     (0.002)           (0.017)
Female/Male wage ratio               -0.0186          -0.314 ***
                                    (0.0714)           (0.101)
Own group size in metro              0.0743           -0.269 ***
                                    (0.0865)           (0.073)
Husband's wage                      0.0002 *            0.0003
                                    (0.0001)           (0.0004)
Husband's experience               -0.0045 ***       -0.0056 ***
                                    (0.0008)           (0.0009)
Husband's years of education         0.0003             0.0006
                                     (0.001)           (0.002)
Husband's veteran status            0.339 ***           0.053
                                     (0.016)           (0.035)
Education dummies                      Yes               Yes
U.S. census region dummies             Yes               Yes
Probability of intermarriage       -0.0269 ***            --
                                    (0.0045)              --
Sex ratio                           0.049 ***             --
                                    (0.0087)              --
N                                    16,520
F-test value of excluded          F-value 51.12   Hansen's   J 15.28
  instruments
Under-identification text          chi square]     p-value   0.001
  [chi square] value              value 144.43

Notes: Not shown here but also included: a third-order
polynominal in experience, veteran status, citizenship,
number of children, ethnicity controls for non-Asians:
black, white, and Hispanic. The dependent variable is log
hourly wage. The explanatory variable of interest is a
binary variable = 1 if husband is U.S.-born and = 0 if
husband is an immigrant. Instruments are the logarithm of
the probability of marrying within the ethnic group sex
ratio in the ethnic group, calculated for the woman's age
group and metropolitan area. Standard errors clustered on
age group-metropolitan area-country of birth cells are shown
in parentheses. Estimates use appropriate person-level
weights from the 2000 U.S. Census.

* Significant at 10%; ** significant at 5%; *** significant at 1%.

Source: Census 2000 5% sample.

TABLE 4

Effects of Intermarriage on Labor Market Employment and
Market Hours

                                                      Employment

                                            Non-Asian       Asian

Raw Estimates (no controls)                0.0644 ***     0.0079 **
                                            (0.0023)      (0.0038)
[R.sup.2]                                    0.0102         0.002
p Value for difference in beta                        0.000
Controls: Educ., Exp., children            0.0115 ***      -0.0022
  Birthplace, Ethnicity, Census Reg.,       (0.0028)      (0.0045)
  Vet. Status, Citizen, YrsUSA. English
[R.sup.2]                                    0.0439        0.0101
p Value for difference in beta                        0.0103
Controls: Row 2 + Spousal Wage, Educ.,     0.0089 ***      -0.0024
  Vet. Status, & Experience                 (0.0029)      (0.0047)
[R.sup.2]                                    0.0776        0.0304
p Value for difference in beta                        0.041
Controls: Row 2 + Row 3 + Metropolitan     0.0078 ***      -0.003
  Characteristics                           (0.0029)      (0.0046)
[R.sup.2]                                    0.0801        0.0307
p Value for difference in beta                        0.0495
N                                            70,836        23,796

                                                      Log Market Hours

                                            Non-Asian       Asian

Raw Estimates (no controls)                -0.0384 ***     -0.0064
                                            (0.0037)      (0.0065)
[R.sup.2]                                    0.0017         0.001
p Value for difference in beta                        0.0000
Controls: Educ., Exp., children            -0.023 ***    -0.019 ***
  Birthplace, Ethnicity, Census Reg.,       (0.0047)      (0.0073)
  Vet. Status, Citizen, YrsUSA. English
[R.sup.2]                                    0.0237         0.025
p Value for difference in beta                        0.6447
Controls: Row 2 + Spousal Wage, Educ.,     -0.0219 ***   -0.0216 ***
  Vet. Status, & Experience                 (0.0044)      (0.0075)
[R.sup.2]                                    0.0733        0.0597
p Value for difference in beta                        0.9771
Controls: Row 2 + Row 3 + Metropolitan     -0.0224 ***   -0.0208 ***
  Characteristics                           (0.0044)      (0.0075)
[R.sup.2]                                    0.0734        0.0601
p Value for difference in beta                        0.8528
N                                            65,915        22,779

Notes: The dependent variable is probability of employment
in the left-hand-side panel, and log weekly market hours in
the right-hand-side panel. The explanatory variable of
interest is a binary variable = 1 if husband is U.S.-born
and = 0 if husband is an immigrant. Estimations use
appropriate person-level weights from the 2000 U.S. census.
Robust standard errors are show in parentheses.

* Significant at 10%; ** significant at 5%;
*** significant at 1%.

Source: Census 2000 5% sample.

TABLE 5

Effects of Intermarriage on Log Hourly Wage, by
Spousal Education

                                                       OLS (Wife's
                           OLS (No Controls)           Controls)

Spouse           Non-Asian       Asian       Non-Asian       Asian
  Education
<High school     0.099 ***     0.205 ***    -0.0327 **      0.0249
                  (0.051)      (0.0146)       (0.051)       (0.015)
N                 19,233         2,043        19,233         2,043
High school      0.078 ***      0.04 *        -0.0059       -0.033
  graduate        (0.011)       (0.025)      (0.0117)       (0.025)
N                 12,917         2,891        12,917         2,891
Some college    0.0603 ***    -0.0465 **    -0.0271 **    -0.0404 **
                  (0.012)       (0.021)       (0.012)       (0.021)
N                 11,565         4,060        11,565         4,060
College          0.055 ***      -0.0048       0.0019        -0.018
  graduate        (0.012)       (0.017)       (0.012)       (0.016)
N                 11,221         7,134        11,221         7,134
College         0.0901 ***     -0.041 **      0.026 *     -0.0614 ***
  graduate +     (0.0173)       (0.021)       (0.016)       (0.022)
N                  6,205         5,467         6,205         5,467

                           OLS (Wife's and
                           Husband's Controls)

Spouse           Non-Asian       Asian
  Education
<High school    -0.0356 **      0.0345
                  (0.053)       (0.015)
N                 19,233         2,043
High school       -0.0008       -0.0328
  graduate        (0.012)      (0.0297)
N                 12,917         2,891
Some college     -0.026 **      -0.0298
                  (0.012)      (0.0216)
N                 11,565         4,060
College           0.0065        0.0008
  graduate       (0.0126)       (0.017)
N                 11,221         7,134
College           0.032 *      -0.0432 *
  graduate +      (0.018)      (0.0233)
N                  6,205         5,467

TABLE 6

Effects of Intermarriage on Log Hours Worked,
by Spousal Education

                   OLS (No Controls)           OLS (Wife Controls)
Spouse
Education        Non-Asian       Asian       Non-Asian       Asian

<High school    -0.0198 **      0.0137        -0.0018       -0.0194
                  (0.01)        (0.027)      (0.0104)       (0.035)
N                 21,588         2,180        25,634         2,574
High school     -0.032 ***     -0.0284 *     -0.022 **      -0.024
                 (0.0077)       (0.016)      (0.0091)       (0.018)
N                 13,741         3,061        14,128         3,042
Some college     -0.0155 *      -0.0105       -0.0129       -0.0096
                 (0.0082)      (0.0136)      (0.0096)      (0.0145)
N                 12,220         4,270        13,868         5,144
College         -0.0346 ***     -0.006      -0.0256 **     -0.027 **
  graduate       (0.0085)       (0.011)      (0.0102)       (0.013)
N                 11,788         7,494        11.183         7,573
College         -0.0296 **       0.022        -0.019        -0.0018
  graduate +     (0.0123)       (0.014)       (0.014)       (0.017)
N                  6,578         5,774         6,785         6,221

                           OLS (Wife and
                           Husband Controls)
Spouse
Education        Non-Asian       Asian

<High school      -0.0007       -0.024
                 (0.0102)       (0.036)
N                 24,952         2,347
High school      -0.019 **      -0.003
                 (0.0088)       (0.019)
N                 12,907         2,677
Some college      -0.0066       -0.008
                 (0.0093)      (0.0144)
N                 12,208         4,580
College         -0.0207 **     -0.0242 *
  graduate       (0.0096)       (0.013)
N                  9,463         6,863
College         -0.0287 **      -0.0014
  graduate +     (0.0135)       (0.016)
N                  5,468         5,640

Notes: The dependent variable is log hourly wage in Table 5,
and log market hours worked in Table 6. Robust standard
errors are shown in parentheses. Spousal education is
divided into five categories. All estimations use
appropriate 2000 Census person weights.

* Significant at 10%; ** significant at 5%; *** significant
at 1%.

Source: Census 2000 5% sample.
COPYRIGHT 2015 Western Economic Association International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2015 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Basu, Sukanya
Publication:Economic Inquiry
Article Type:Abstract
Date:Oct 1, 2015
Words:11407
Previous Article:Policy variation, labor supply elasticities, and a structural model of retirement.
Next Article:Voluntary activities and daily happiness in the United States.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |