An American Dream Unfulfilled: The Limited Mobility of Mexican Americans*

An American Dream Unfulfilled: The Limited Mobility of Mexican Americans* Gretchen Livingston, University of Pennsylvania Joan R. Kahn, University of Maryland at College Park Objective. We build on past research regarding immigrant group adaptation by examining the wages of first-, second-, and third-generation Mexican-American men and women and empirically evaluating if past theories of immigrant incorporation apply to the Mexican-American case. Methods. We use the 1989 Latino National Political Study and the 1990/1991 Panel Studies of Income Dynamics and OLS regressions to estimate the effects of generation and human capital on wages. Results. Immigrant men and women report lower wages than their second- and thirdgeneration counterparts, but once human capital controls are added, the wage pattern becomes one of steady decline across generations for men, and stagnation or marginal decline across generations for women. Conclusions. Our results generally contest the applicability of linear assimilation hypotheses to the Mexican-American experience, while lending some credence to the selectivity and immigrant optimism hypotheses. Results also indicate the importance of developing more contextualized immigrant adaptation frameworks. Background Mexican-origin persons comprise the majority of Hispanics living in the United States and their presence is expected to increase drastically in the coming decades (Fix and Passel, 1994:26; Edmonston and Passel, 1994; del Pinal and Singer, 1997:7). Despite their long history in the United States, Mexican-origin peoples continue to perform as badly as, or worse than, other American minorities on many measures of well-being. Poor educational achievement and occupational placement translate into low wages and high poverty rates rivaling those of African Americans (Trejo, 1997; del Pinal and Singer, 1997:38). Given projected increases in the population of Mexican Americans, and their chronically low socioeconomic standing, it *Direct all correspondence to Gretchen Livingston, Population Studies Center, University of Pennsylvania, 3718 Locust Walk, Philadelphia, Pennsylvania 19104 <glivings@ pop.upenn.edu>. The data used in this research is available through the Inter-University Consortium for Political and Social Research at the University of Michigan. Details of data coding available upon request from the first author. We thank Paula England, Grace Kao, Suzanne Bianchi, Reeve Vanneman, and Doug Massey for their helpful comments. SOCIAL SCIENCE QUARTERLY, Volume 83, Number 4, December 2002 2002 by the Southwestern Social Science Association

1004 Social Science Quarterly seems essential to develop a framework to characterize the way they incorporate into the U.S. economy. Unfortunately, research regarding the mobility of Mexican Americans is limited. It has been hindered primarily by inadequate data and methods that have contributed to, and been exacerbated by, underdeveloped theoretical frameworks. The bulk of immigrant incorporation research examines mobility as a consequence of nativity (i.e., foreign born vs. native born), not of generation (e.g., first generation, second generation, third generation). Given the curvilinear pattern of cross-generational mobility proposed by some theories of immigrant incorporation, the use of nativity is a questionable means of measuring immigrant group mobility, which precludes explicit theoretical evaluation. Furthermore, much of the research that has been conducted is based on cross-sectional comparisons of aggregated immigrant/ethnic groups, which can produce spurious results by conflating cross-generational changes with compositional changes in the national origins of immigrants (Borjas, 1994). Also, since many studies are based on data from the 1960s and 1970s, and focus on male labor force performance, they do not reflect the more recent bifurcation of the labor market or the feminization of the immigrant and Mexican-origin labor force in the United States (Danziger and Gottschalk, 1995:14; Levy, 1994; Pedraza, 1991; McKeever and Klineberg, 1999; Allensworth, 1997; Kahn and Whittington, 1996). By using the Panel Study of Income Dynamics/Latino National Political Survey (PSID/LNPS), we overcome these shortcomings. We examine the economic incorporation of Mexican-American men and women by reviewing cross-generational differences in economic achievement for the first, second, and third generations. Additionally, we synthesize these findings in order to provide a preliminary evaluation of the applicability of several theories of immigrant incorporation to the Mexican-American experience. Theory and Evidence Regarding Economic Incorporation Several theories regarding the cross-generational pattern of economic incorporation among ethnic groups have been proposed. The classic assimilation and modified assimilation hypotheses both imply that immigrant group incorporation will increase across successive generations as groups adjust to the American way of life. Whereas the former hypothesis emphasizes the role of individual factors in immigrant incorporation, the latter acknowledges that discrimination may hinder the ability of immigrants to fully incorporate into the destination country (Gordon, 1964; Bean et al., 1994; Portes and Zhou, 1992; Hirschman, 1996). The selectivity and immigrant optimism hypotheses both predict a curvilinear pattern of incorporation, characterized by relatively high second-generation achievement. They argue that immigrant wages are low because, despite being positively selected on motivation,

Mexican-American Mobility 1005 immigrants often lack U.S.-specific human capital. Second-generation persons typically earn relatively good wages because they inherit a great deal of motivation from their parents and benefit from high levels of human capital. The third generation earns lower wages on average because, although it possesses human capital similar to the second generation, it does not necessarily acquire the same high motivation level. Whereas the selectivity hypothesis emphasizes the role of individual differences in affecting motivation, and thus wage outcomes, the immigrant optimism hypothesis suggests that incorporation patterns are due to structural constraints and discrimination, which discourage the motivation to achieve among later generations (Kao and Tienda, 1995; Bean et al., 1994; Carliner, 1980; Chiswick, 1977, 1978). Although they vary in the emphasis placed upon it, all of these theories acknowledge that human capital plays a role in intergenerational mobility. Years of formal education and work experience are considered important predictors of worker productivity and wage, particularly if obtained in the host country. English proficiency, too, presumably benefits productivity and wage. Empirical work confirms the association of these human capital factors with labor force outcomes and generation or nativity, but it has not consistently substantiated any one hypothesis of immigrant incorporation (Kahn and Whittington, 1996; McKeever and Klineberg, 1999; Bean and Tienda, 1987; Long, 1980; Carliner, 1980; Stolzenberg and Tienda, 1997; Chiswick, 1977, 1978; Allensworth, 1997; Trejo, 1997; Padilla and Glick, 2000). With few exceptions (Long, 1980), wage and earnings growth has been observed between the foreign born and native born, but evidence regarding differences between the second and third generations is limited (Bean and Tienda, 1987:284; Chiswick, 1977, 1978; Allensworth, 1997; Carliner, 1980; Borjas, 1993, 1996; Stier and Tienda, 1992; Padilla and Glick, 2000). We know of no research that explicitly considers this issue for Mexican-origin women, and the little research regarding Mexican-origin men reports either increasing economic achievement across three generations (Borjas, 1993), or wage stagnation beyond the second generation (Trejo, 1997). Analytic Approach Our research builds on the past literature by employing relatively recent data to compare the wages of first-, second-, and third-generation Mexican- American men and women and to consider the determinants of any differences that emerge. We also develop a simple framework to evaluate the applicability of the four immigrant adaptation theories to the Mexican- American population.

1006 Social Science Quarterly A monotonic wage increase across generations lends support to the classic assimilation and modified assimilation hypotheses. These theories imply that controlling for human capital will diminish generational differences in economic achievement, particularly between native-born groups (incomplete assimilation may prevent immigrants as a group from ever fully incorporating). A curvilinear pattern of wage mobility (where the second generation enjoys the highest relative wage) is consistent with the selectivity and immigrant optimism hypotheses. Both of these theories suggest that human capital rises across all generations, while motivation to achieve declines beyond the second. Consequently, they predict that controlling for human capital will essentially eliminate wage differences between immigrants and the second generation, while creating a relative wage deficit for the third generation. 1 Data and Methods The analysis is based on data from the 1990 and 1991 Panel Studies of Income Dynamics, along with the 1989 Latino National Political Survey. We include those self-identified Mexican-origin persons who participated in the 1989 LNPS and the 1990 and 1991 PSID. 2 The final sample includes 553 respondents (288 men, 265 women). Our univariate statistics are weighted by a factor that generates an LNPS/PSID sample representative of Mexican-origin persons in the United States (Duncan et al., 1995). However, we do not weight our regressions or include a variable to correct for the selectivity of including only labor force participants in our analyses (Berk, 1983; Heckman, 1979) because doing so reduces the efficiency of the models and may exacerbate any effects of selectivity bias (Winship and Radbill, 1994; DuMouchel and Duncan, 1983; Stolzenberg and Relles, 1990, 1997; Winship and Mare, 1992). Omitting weights and the Heckman correction factor does not produce significant changes in the pattern or significance of any regression coefficients. We base our outcome measure on hourly wage. To diminish wage variation, we calculate an average hourly wage for 1989 and 1990 and take its natural log. We try to reduce wage volatility not only to gain a more accurate representation of the patterns of wage across generations, but also to accentuate any wage differences that do exist; this is particularly important 1 An obvious next step in this analysis would be to include controls for local labor market structure and discrimination, since the immigrant optimism and modified assimilation hypotheses both point to the importance of these factors in affecting achievement. Due to data limitations, however, we were unable to do so. 2 Due to data limitations, only persons who were the head of household, or the spouse/ significant other of the head of household, are included. In an effort to exclude students, retirees, and workers compensated by in-kind benefits, we limit our analysis to labor force participants aged 25 to 64 who report earning average hourly wages of $1 or more.

Mexican-American Mobility 1007 TABLE 1 Sample Characteristics by Sex and Generation First Generation Second Generation Third Generation Men Hourly wage 7.84 (1.68) 10.42 (1.64) 9.90 (1.72) Education 7.45 (4.36) 12.00 (3.15) 12.49 (2.34) English proficiency 2.54 (1.14) 4.98 (1.25) 5.58 (1.01) U.S. work experience 15.70 (9.54) 21.37 (11.65) 18.04 (7.74) N 165 54 69 Women Hourly wage 5.00 (1.64) 6.63 (1.74) 6.40 (1.98) Education 7.34 (3.91) 10.86 (3.44) 11.36 (3.22) English proficiency 2.46 (1.19) 4.98 (1.06) 5.20 (1.12) U.S. work experience 11.29 (7.52) 18.14 (10.10) 15.80 (9.73) N 112 59 94 NOTE: Standard deviations in parentheses. for women, who tend to experience less wage elasticity than men (Bean and Tienda, 1987; Stier and Tienda, 1992). The primary independent variable is measured by a series of dummies representing generation. Respondents born abroad who migrated after age five are classified as first generation, respondents who were born in the United States or moved here prior to age six and who report at least one Mexican-born parent, are classified as second generation, and respondents who were born in the United States to two U.S.-born parents are categorized as members of the third generation (though in fact this category includes some persons from higher-order generations as well). We also include variables to control for human capital. Education is based on years of formal schooling. English proficiency is determined via a self-rated score measured on a seven-point scale ranging from Speaks no English to Speaks no Spanish. U.S. work experience equals the total years of work experience since age 18. For persons who migrated to the United States after age 18, we limit their possible U.S. work experience to the number of years since migrating.

1008 Social Science Quarterly TABLE 2 Annotated Results for OLS Regressions on Log-Wage Model 1 Generation Controls Men First generation 0.2850 2,3 (0.0815) Model 2 Model 1+ Human Capital Controls 0.0997 3 (0.0997) Second generation Third generation 0.0516 1 (0.0944) 0.1478 1 (0.0899) R 2 0.0575 0.2171 Women First generation 0.2812 2,3 (0.0933) 0.1355 (0.1135) Second generation Third generation 0.0340 1 (0.0963) 0.0574 (0.0869) R 2 0.0478 0.2475 NOTE: Standard errors in parentheses. Significantly different (p < 0.05) than: 1 First generation, 2 Second generation, 3 Third generation. Model 2 includes controls for years of education, English proficiency, years of U.S. work experience, and years of U.S. work experience squared. Results Univariate analyses in Table 1 confirm that education and English proficiency increase markedly across generations for men and women, with the most dramatic increases occurring between the first and second generations. Among both men and women, U.S. work experience is lowest for immigrants and highest among the second generation. Experience drops among third-generation Mexican Americans, due primarily to the fact that they are younger on average than their second-generation counterparts. In the absence of controls, the cross-generational wage pattern for men is curvilinear. First-generation men report an hourly wage of $7.84, second-generation men report a wage of $10.42, and third-generation men report an average wage of $9.90. Wage levels are markedly lower among women, although the cross-generational pattern is similar. First-generation women report a wage of $5.00 on average, the second-generation wage is about $6.63, and the third-generation wage is $6.40. In Table 2 we include portions of two regression models predicting logwage for men and women. We present the generation coefficients for models where second generation is the omitted category, so each coefficient can be interpreted as the percent difference in wage compared to the second generation. Although we only report the coefficients produced by using the

Mexican-American Mobility 1009 second generation as the omitted category, we tested for statistically significant differences between all generation coefficients and indicate those distinctions as well. Model 1 controls for generation only; Model 2 adds controls for education, U.S. work experience, U.S. work experience squared, and English proficiency. 3 The results in Model 1 of Table 2 echo our univariate findings that the cross-generational wage pattern among both men and women is nonlinear, with immigrants earning significantly less than both native-born groups. Wages of second-generation men are about 28 percent higher than those of immigrant men, and 5 percent higher than those of third-generation men. Among women, the second generation earns wages that are 28 percent higher than immigrant wages, and 3 percent higher than third-generation wages. These findings provide some support for the selectivity and immigrant optimism hypotheses, while challenging the applicability of both the classic assimilation and modified assimilation theories, which predict monotonic wage increases across generations in the absence of controls. The selectivity and immigrant optimism hypotheses suggest that crossgenerational wage patterns are influenced by declines in motivation and increases in human capital across generations. We have no measure of motivation, but the theories predict that holding human capital constant will diminish the wage differences between immigrants and the second generation, and magnify the relative wage deficit of the third generation. Indeed, this pattern emerges in the top panel of Model 2. The significant wage difference between immigrants and the second generation disappears, while the third-generation coefficient declines precipitously. Without the benefit of their high human capital, third-generation men would earn about 15 percent less than their second-generation counterparts and 25 percent less than comparable immigrant men. The story for women is a bit less clear. The fact that wage differences between generations are not highly significant once human capital is controlled for may suggest that human capital differences are driving generational variations in female wage; this would correspond most closely with the classic or modified assimilation hypothesis. However, it is the case that immigrant women earn wages that are marginally (p < 0.10) higher than comparable third-generation women, which suggests that the selectivity or immigrant optimism hypotheses could be applicable. Conclusions We find that for Mexican-origin men and women, wage growth occurs between the first generation and both of the native-born groups, but no 3 In earlier iterations of this work, we included models controlling for family structure as well, but doing so produced no notable change in the regression coefficients or the explanatory power of Model 2, so we exclude them from this analysis.

1010 Social Science Quarterly notable growth occurs between the second and third generations. In fact, the third generation experiences a slight wage decline when compared to the second generation. Controlling for human capital eradicates the firstgeneration wage deficits for men and women, thus confirming that low immigrant wages are primarily due to low levels of U.S.-relevant human capital. Conversely, the relative wage deficit of the third generation increases once human capital is controlled for; so were it not for their relatively high levels of human capital, third-generation Mexican Americans would perform even more poorly in the U.S. labor market, as compared to more recent generations. Of course, a lack of wage growth across generations does not necessarily imply a lack of economic adaptation or well-being (Alba and Nee, 1997). The relatively high wages of some respondents, for instance, could be considered indicative of the fact that Mexican Americans have already made it. In reality, though, the average wages of third-generation Mexican- American respondents still fall more than 25 percent below the wages of comparable non-hispanic whites. Although we argue that the mobility of Mexican Americans continues to be limited, the forecast for the future need not be pessimistic. Like Trejo (1997), we find that that differences in human capital endowments, not wage structure, account for most of the cross-generational wage patterns that presently exist. To some extent, this is a good sign it reaffirms the commonsense notion that increasing human capital should lead to improvements in the labor force outcomes of Mexican Americans. In addition to the policy implications, our finding of nonlinear wage mobility generally refutes linear assimilation theories and lends some support to the selectivity and immigrant optimism hypotheses. However, the strength of our conclusions varies somewhat by gender. Although this could simply be due to gender differences in wage elasticity, we think it points to the importance of considering more nuanced conceptualizations of immigrant adaptation, such as the segmented assimilation hypothesis (Portes, 1997; Portes and Rumbaut, 1996; Portes and Zhou, 1992), which allows for the fact that even within one migrant group, experiences of adaptation can vary markedly. REFERENCES Alba, Richard, and Victor Nee. 1997. Rethinking Assimilation Theory for a New Era of Immigration. International Migration Review 31:826 74. Allensworth, Elaine M. 1997. Earnings Mobility of First and 1.5 Generation Mexican- Origin Women and Men: A Comparison with U.S.-Born Mexican Americans and Non- Hispanic Whites. International Migration Review 31:386 410. Bean, Frank D., Jorge Chapa, Ruth R. Berg, and Kathryn A. Sowards. 1994. Educational and Sociodemographic Incorporation Among Hispanic Immigrants to the United States. In Jeffrey Passel and Barry Edmonston, eds., Immigration and Ethnicity. Washington, D.C.: Urban Institute.

Mexican-American Mobility 1011 Bean, Frank D., and Marta Tienda. 1987. The Hispanic Population of the United States. New York: Russell Sage Foundation. Berk, Richard A. 1983. An Introduction to Sample Selection Bias in Sociological Data. American Sociological Review 48:386 98. Borjas, George J. 1993. The Intergenerational Mobility of Immigrants. Journal of Labor Economics 11:113 35.. 1994. The Economics of Immigration. Journal of Economic Literature 32:1667 1717.. 1996. The Earnings of Mexican Immigrants in the United States. Journal of Development Economics 51:69 98. Carliner, Geoffrey. 1980. Wages, Earnings, and Hours of First, Second and Third Generation American Males. Economic Inquiry 18:87 102. Chiswick, Barry R. 1977. Sons of Immigrants: Are They at an Earnings Disadvantage? American Economic Review 67:376 80.. 1978. The Effect of Americanization on the Earnings of Foreign-Born Men. Journal of Political Economy 86:897 921. Danziger, Sheldon, and Peter Gottschalk. 1995. America Unequal. New York: Russell Sage. Del Pinal, Jorge, and Audrey Singer. 1997. Generations of Diversity: Latinos in the United States. Washington, D.C.: Population Reference Bureau. DuMouchel, William H., and Greg J. Duncan. 1983. Using Sample Survey Weights in Multiple Regression Analyses of Stratified Samples. Journal of the American Statistical Association 78:535 43. Duncan, Greg, Martha Hill, James Morgan, and Sandra Hofferth. 1995. A Panel Study of Income Dynamics: Procedures and Tape Codes, 1990 Interviewing Year. Ann Arbor, Mich.: University of Michigan Survey Research Center. Edmonston, Barry, and Jeffrey S. Passel. 1994. The Future Immigrant Population of the United States. In Jeffrey Passel and Barry Edmonston, eds., Immigration and Ethnicity. Washington, D.C.: Urban Institute. Fix, Michael, and Jeffrey S. Passel. 1994. Immigration and Immigrants: Setting the Record Straight. Washington, D.C.: Urban Institute. Gordon, Milton M. 1964. Assimilation in American Life. New York: Oxford University Press. Heckman, James J. 1979. Sample Selection Bias as a Specification Error. Econometrica 47:153 61. Hirschman, Charles. 1996. Studying Immigrant Adaptation from the 1990 Population Census: From Generational Comparisons to the Process of Becoming American. In Alejandro Portes, ed., The New Second Generation. New York: Russell Sage Foundation. Kahn, Joan R., and Leslie A. Whittington. 1996. The Labor Supply of Latinas in the U.S.A.: Comparing Labor Force Participation, Wages, and Hours Worked with Anglo and Black Women. Population Research and Policy Review 15:45 73. Kao, Grace, and Marta Tienda. 1995. Optimism and Achievement: The Educational Performance of Immigrant Youth. Social Science Quarterly 76:1 19. Levy, Frank. 1994. Incomes and Income Inequality. In Reynolds Farley, ed., State of the Union: Economic Trends. New York: Russell Sage Foundation.

1012 Social Science Quarterly Long, James E. 1980. The Effect of Americanization on Earnings: Some Evidence for Women. Journal of Political Economy 88:620 29. McKeever, Matthew, and Stephen L. Klineberg. 1999. Generational Differences in Attitudes and Socioeconomic Status Among Hispanics in Houston. Sociological Inquiry 69:33 50. Padilla, Yolanda C., and Jennifer E. Glick. 2000. Variations in the Economic Integration of Immigrant and U.S.-Born Mexicans. Hispanic Journal of Behavioral Sciences 22:179 93. Pedraza, Silvia. 1991. Women and Migration: The Social Consequences of Gender. Annual Review of Sociology 17:303 25. Portes, Alejandro. 1997. Immigration Theory for a New Century: Some Problems and Opportunities. International Migration Review 31:799 825. Portes, Alejandro, and Ruben Rumbaut. 1996. Immigrant America: A Portrait. Berkeley, Cal.: University of California. Portes, Alejandro, and Min Zhou. 1992. Gaining the Upper Hand: Economic Mobility Among Immigrant and Domestic Minorities. Ethnic and Racial Studies 15(4):491 522. Stier, Haya, and Marta Tienda. 1992. Family, Work and Women: The Labor Supply of Hispanic Immigrant Wives. International Migration Review 26:1291 1313. Stolzenberg, Ross M., and Daniel A. Relles. 1990. Theory Testing in a World of Constrained Research Design. Sociological Methods and Research 18:395 415.. 1997. Tools for Intuition About Sample Selection Bias and Its Correction. American Sociological Review 62:494 507. Stolzenberg, Ross M., and Marta Tienda. 1997. English Proficiency, Education, and the Conditional Assimilation of Hispanic and Asian Origin Men. Social Science Research 26:25 51. Trejo, Stephen J. 1997. Why Do Mexican Americans Earn Low Wages. Journal of Political Economy 105:1235 68. Winship, Christopher, and Robert D. Mare. 1992. Models for Sample Selection Bias. Annual Review of Sociology 18:325 50. Winship, Christopher, and Larry Radbill. 1994. Sampling Weights and Regression Analysis. Sociological Methods and Research 23:230 57.