Second-Generation Decline or Advantage? Latino Assimilation in the Aftermath of the Great Recession 1

Transcription

1 Second-Generation Decline or Advantage? Latino Assimilation in the Aftermath of the Great Recession 1 Van C. Tran Columbia University Nicol M. Valdez Columbia University This article addresses the debate on second-generation advantage and decline among Latinos by providing a post-recession snapshot based on geocoded data from the Current Population Survey ( ). It reports three findings. First, second-generation Mexicans and Puerto Ricans are at a disadvantage, whereas other Latinos have achieved 1 We are grateful to Jody Agius-Vallejo, Rene Flores, Norma Fuentes-Mayorga, Herb Gans, David Goodwin, David Grusky, Tomas Jimenez, Phil Kasinitz, Doug Massey, Beth Mattingly, Edward Telles, Charles Varner, Mary Waters, Chris Winship, and participants of the Stanford Center for Poverty and Inequality s Hispanic Mobility and Inequality Working Group for helpful comments on previous drafts. This research has also been presented at Barnard College s Forum on Migration, the Department of Sociology at Hunter College of the City University of New York, Harvard University s Migration and Immigrant Incorporation Workshop, Yale University s Center for Research on Inequalities and the Life Course, the University of Wisconsin s Institute for Research on Poverty, the US Department of Health and Human Services Administration for Children and Families (ACF), the 2013 APPAM Annual Meeting, and the 16th Annual Welfare Research and Evaluation Conference. We thank our diverse audiences at these venues as well as the IMR editor and reviewers for their constructive feedback which has strengthened our research. This research was partially supported by a Poverty Center Grant awarded to the Center on Poverty and Inequality at Stanford University (grant number AE00101) from the US Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation, and awarded by the Substance Abuse Mental Health Service Administration and a sub-award (grant number H79 AE S1) titled Poverty, Inequality, and Mobility among Hispanic Populations: An Innovative Subgrant Research Program at the Stanford Center on Poverty and Inequality funded by ACF s Office of Planning, Research, and Evaluation. All remaining errors are our own. Please direct all correspondence to Van C. Tran, Department of Sociology, Columbia University at vantran@columbia.edu by the Center for Migration Studies of New York. All rights reserved. DOI: /imre IMR Volume ** Number ** (Fall 2015):1 36 1

2 2 INTERNATIONAL MIGRATION REVIEW parity with native majority peers. Second, second-generation Latinos report significant progress compared to their parents and there is no evidence of a second-generation decline. Third, there is no difference in outcomes among second-generation Mexicans by immigrant destination type. Overall, these analyses yield an optimistic assessment of second-generation progress, while noting potential stagnation among third- and higher-generation Mexicans. INTRODUCTION Over the last two decades, four key trends have transformed the Latino experience in the United States. First, the Latino population emerged as the largest minority group in the country, comprising 16.3 percent of the US population, or 50.5 million people, in This represents a growth of 43 percent from 2000 to The Latino population is projected to continue to increase in the coming decades, reaching million people or 30 percent of the US population in 2050 (US Census Bureau 2011). Second, members of the Latino second generation (i.e., those who were born in the United States to foreign-born parents) are coming of age in sizable numbers. As a result, understanding their socioeconomic attainment in young adulthood provides useful clues to how their presence will transform patterns of ethnic and racial inequality in American society in the coming decades (Portes and Rumbaut 2001; Alba and Nee 2003; Kasinitz et al. 2008; Lee and Bean 2010). Third, the constant replenishment of the Latino population with new immigrants has increased the heterogeneity among Latinos by ethnic origin, immigrant generation, social class background, and legal status (Tienda and Mitchell 2006; Telles and Ortiz 2008; Alba, Jimenez, and Marrow 2014; Waters 2014). Lastly, Latinos are increasingly settling in smaller cities and towns in new immigrant destinations outside of traditional immigrant gateways (Massey 2008; Marrow 2011; Kritz and Gurak 2015). This profound shift is both an unintended consequence of stricter border control policies that occurred in the 1990s (Massey, Durand, and Malone 2002) and a direct result of demographic changes and economic restructuring in these new destinations (Marrow 2011; Flippen and Parrado 2012). Latino assimilation thus takes on both regional and national relevance. In light of these trends, this article addresses three sets of related questions. First, how do second-generation Latinos fare in comparison

3 SECOND-GENERATION DECLINE OR ADVANTAGE? 3 to their native peers? Our strategy is to compare the socioeconomic attainment among the Latino second generation to that of their peers from the native majority group (non-hispanic white) and native minority groups (non-hispanic black and Puerto Rican). Altogether, these three groups provide different contextual benchmarks to evaluate outcomes among the second generation (Park, Myers, and Jimenez 2014). Second, how do second-generation Latinos fare in comparison to their parents? Because the Current Population Survey (CPS) does not contain data on the respondents parents, our strategy is to compare second-generation Latinos to their proxy first-generation parents, using the lagged birth cohort method (Farley and Alba 2002). Third, how do second-generation Latinos in new destinations fare in comparison to those in traditional gateways? Our strategy is to compare their attainment to that of corresponding Latino and non-latino peers across destination types. Theoretically, we argue that the choice of the reference group to which we compare Latino progress and attainment is rather consequential to the conclusion we draw from our research. We seek to bring clarity to this debate by distinguishing between two subsets of comparisons. On the one hand, debates on second-generation advantage or disadvantage focus on comparisons between second-generation individuals and their nativeborn peers. On the other hand, debates on second-generation decline or progress focus on comparisons between second-generation individuals and their first-generation parents. In addition, we highlight a third set of comparisons on second-generation outcomes between old and new immigrant destinations. Our analyses report three main findings. First, we find that second-generation Mexicans and Puerto Ricans are at a disadvantage in comparison to the native minority groups, whereas many Latino groups have achieved parity with their native majority peers. Second, there is no evidence of a second-generation decline and the Latino second generation reports significant progress compared to their first-generation parents. Third, there is no difference in outcomes among second-generation Mexicans by immigrant destination type. These are not new questions, but prior studies were based on prerecession data and the most recent evidence is now almost a decade old. A post-recession evaluation of outcomes among the Latino second generation with the most recent evidence is not only timely, but also relevant because recent shifts in our politics and our economy have resulted in a rather negative context of reception for Latinos in the post-recession era. Specifically, we point to three key factors. First, the great recession

4 4 INTERNATIONAL MIGRATION REVIEW disproportionately affected the Latino population. For example, the number of Latino children in poverty rose sharply from 4.4 million in 2007 to 6.1 million in This was a 36.3 percent increase, compared to 17.6 percent for non-latino whites and 11.7 percent for non-latino blacks during the same period (Lopez and Velasco 2011). Another direct consequence of the great recession was significant wealth loss, especially among Latinos. According to one study from the Urban Institute, Latinos lost 44 percent of their overall wealth between 2007 and 2011, compared to only 11 percent lost among non-hispanic whites and 31 percent lost among non-latino blacks. The majority of this lost wealth was concentrated in home equity as many Latinos were affected by the foreclosure crisis (McKernan et al. 2013). Second, the debate around the undocumented population (estimated at 12.2 million in 2007 and 11.2 million in 2011) remains highly contentious, especially so in the aftermath of the great recession, where millions of Americans were out of work. As a result, Hispanics, and especially Mexicans, were not only stigmatized, but also bore the brunt of this recent wave of anti-immigrant sentiment. During this period, the number of deportations also reached a record high, rising from 359,795 in 2008 to 438,421 in 2013, an increase of 21.9 percent over five years (US Department of Homeland Security 2013). This debate about the undocumented also unfolded alongside the DREAMers movement, which transformed the national debate on immigrant rights and brought about significant immigrant reforms (Nicholls 2013). Third, Latinos are increasingly settling in new destinations, where they face significant prejudice from both non-hispanic blacks and non-hispanic whites, as they do not fit neatly into the pre-existing binary racial order in rural and small-town America (Marrow 2011). Substantively, this article contributes to previous research in four ways. First, it will provide not only the most recent trends, but also one of the first post-recession snapshots of socioeconomic assimilation among Latinos. Second, whereas previous research focused on specific immigrant regions (Portes and Rumbaut 2001; Kasinitz et al. 2008; Telles and Ortiz 2008; Park, Myers, and Jimenez 2014) or on one ethnic group such as Mexicans (Perlmann 2005; Telles and Ortiz 2008; Jimenez 2010; Agius- Vallejo 2012; Park, Myers, and Jimenez 2014), this article provides one of the first national overviews of second-generation Latinos in young adulthood from ten ethnic groups. Third, it will evaluate competing theoretical predictions about the possibility of a second-generation advantage and disadvantage (Portes and Zhou 1993; Portes and Rumbaut 2001;

5 SECOND-GENERATION DECLINE OR ADVANTAGE? 5 Kasinitz et al. 2008; Alba, Kasinitz, and Waters 2011) as well as a second-generation decline and progress (Gans 1992; Smith 2003; Telles and Ortiz 2008). More broadly, it re-examines assumptions about Latino assimilation and shows how the choice of a reference group to benchmark progress shapes conclusions on Latino mobility. Finally, we examine socioeconomic assimilation among the second generation in new immigrant destinations. Although the existing literature on new immigrant destinations has primarily addressed the first generation (Molly and Muller 2012; Fischer 2010; Stamps and Bohon 2006), we know relatively little about the second generation, given their recent presence there (but see Morando 2013). ASSIMILATION AND RACIALIZATION OF LATINOS Over the last two decades, much of scholarly research and debate have focused on the assimilation of post-1965 immigrants and their children (Kasinitz et al. 2008; Alba and Nee 2003; Portes and Rumbaut 2001; Smith 2006; Waters and Jimenez 2005). Two competing theoretical perspectives inform the contentious debate on the new second generation: the theory of straight-line assimilation (and its variants) and the theory of segmented assimilation (Alba, Kasinitz, and Waters 2011; Haller, Portes, and Lynch 2011). Originating from the experience of European immigrants at the turn of the twentieth century, the theory of straight-line assimilation (and its variants) posits that progressive assimilation is linked to socioeconomic advancement across successive immigrant generations. Assimilation is the final stage of adjustment that is achieved when immigrant offspring manage to attain a socioeconomic level on par with native-born whites. For the immigrants who arrived at the turn of the twentieth century, upward mobility not only occurred simultaneously with their increasing assimilation into American society, but also was the master trend for European ethnic groups (Alba and Nee 2003). In contrast, scholars raised the possibility of downward mobility among the post-1965 second generation (Gans 1992; Portes and Zhou 1993). Specifically, the theory of segmented assimilation has pointed to the macro-structural sources of vulnerability, especially among groups with darker phenotypes: their racial distinctiveness, the poor quality of urban schools, their concentration in poor neighborhoods in close proximity to African Americans, an increasingly bifurcated economy with fewer well-paid jobs, and racial discrimination.

6 6 INTERNATIONAL MIGRATION REVIEW Latinos are at the forefront of this debate for four reasons. First, they comprise the majority of the immigrant population and their children. Second, scholars note their relatively lower levels of socioeconomic attainment with some predicting that Mexicans are most at risk of downward mobility into the new rainbow underclass (Portes and Rumbaut 2001, 45). 2 Third, Latinos occupy an in-between racial status in the American binary racial structure, making it all the more important to explore how they might fit into the American color line (Marrow 2011; Roth 2012). Finally, Latinos and especially Mexicans comprise the majority of the undocumented population (Passel and Cohn 2011) and the lack of legal status negatively affects second-generation attainment (Bean 2011; Greenman and Hall 2013). Another reason for Latinos to feature prominently in this debate is their experiences with racialization, especially for ethnic groups whose members have significant indigenous or African components, such as Mexicans, Puerto Ricans, and Dominicans (Massey 2008; Telles and Ortiz 2008; Vaquera, Aranda, and Gonzales 2014). Citing these ethnic groups low level of human capital, darker skin tones, and lack of legal status, this body of work suggests that the persistence of racial discrimination might prevent them from being fully assimilated into American society. This is especially the case for Mexicans given their demographic dominance, their history of circular migration, and their significant internal heterogeneity and constant replenishment (Jimenez 2010; Alba, Jimenez, and Marrow 2014). For example, it is not uncommon to hear the popular refrain that if one takes out the Mexicans, there will be no evidence for segmented assimilation among immigration researchers (Portes and Rumbaut 2001, 276). For these reasons, Mexicans have come to be seen as an exemplary case of segmented assimilation (Jimenez 2010, 16). Empirical evidence to date remains mixed for Latinos, spanning the spectrum of racialization on the one hand and full integration on the other (Alba, Jimenez, and Marrow 2014). On the one hand, Portes and Rumbaut (2001, 277) concluded that the Mexican second-generation was the textbook example of a major group that is most at risk of downward assimilation. In an earlier analysis using nationally representative data, Farley and Alba (2002) showed that Mexicans and Puerto Ricans reported the worst socioeconomic outcomes among second-generation groups. In 2 In contrast, others note their propensity for upward mobility by drawing on the minority culture of mobility that might be available to them within their ethno-racial groups (Neckerman, Carter, and Lee 1999; Agius-Vallejo 2012).

7 SECOND-GENERATION DECLINE OR ADVANTAGE? 7 their analysis of data from the second generation in New York, Kasinitz and others (2008) showed that second-generation Dominicans and South Americans had an advantage over Puerto Ricans in young adulthood, although this data did not include Mexicans. On the other hand, several recent studies found little evidence of downward mobility (Waldinger and Feliciano 2004; Waldinger 2007; Luthra and Waldinger 2010; Agius-Vallejo 2012; Rendon 2014). Drawing on unique longitudinal data across four generations, Telles and Ortiz (2008) provided an important perspective on this debate by showing clear progress between first- and secondgeneration Mexicans, although Telles and Ortiz also showed clear stagnation in the third and fourth generations, providing support for the racialization perspective. What might account for such different speculations on and conclusions about Latino second-generation outcomes? First, different studies have captured the Latino second generation in different life stages: from early adolescence to young adulthood (Portes and Rumbaut 2001; Kasinitz et al. 2008; Telles and Ortiz 2008). For example, the high aspirations for educational success that are universal among many adolescents from immigrant background might not be fully realized in young adulthood because of the reality of low-performing public schools and segregated neighborhoods that affect many Latinos (Rendon 2014). Second, different studies partially capture both the anxiety and optimism that are not only emblematic of public sentiments toward immigration and reflect the economic conditions of the period. For example, it is no coincidence that Portes and Zhou (1993) and Gans (1992) warned about second-generation decline in the 1990s during a serious recession while Kasinitz and others (2008) documented mostly optimistic conclusions, in part because data were collected by the latter group during the Clinton era s economic boom. Third, the majority of prior studies have focused on traditional gateways such as California, New York, and Texas (Portes and Rumbaut 2001; Kasinitz et al. 2008; Telles and Ortiz 2008; Park, Myers, and Jimenez 2014), while only a few have begun to focus on new immigrant destinations such as North Carolina (Marrow 2011) and Georgia (Morando 2013). ASSESSING INTERGENERATIONAL PROGRESS AMONG LATINOS Given the racial stratification of American society, the choice of the reference group can shape our assessment of intergenerational progress and the

8 8 INTERNATIONAL MIGRATION REVIEW conclusions drawn from the research. As Kasinitz and others (2008, 15) have observed: Attempting to measure the progress or the assimilation of the second generation immediately poses questions: Assimilation into what? And progress compared to whom? A century ago, arriving European immigrants were integrated into an American society that was predominantly white in racial composition, with African Americans in a position of relative disadvantage. Today, Latino immigrants and their children encounter an American society that is diverse in both race and social class. Recent research has increasingly paid attention to the choice of the reference group in benchmarking second-generation progress (Kasinitz et al. 2008; Telles and Ortiz 2008; Park, Myers, and Jimenez 2014). We build on this body of work by distinguishing between two sets of comparisons in this paper: (1) comparisons with native peers, and (2) comparisons with parents. The first set of comparisons between the second generation and their native-born peer informs debates on second-generation advantage or disadvantage (Portes and Rumbaut 2001; Alba and Nee 2003; Kasinitz et al. 2008). In such comparisons, there are at least five possible outcomes, depending on the choice of either the native majority group (e.g., non-hispanic white) or the native minority groups (e.g., non-hispanic black and Puerto Rican) as the reference group. First, the second generation outperforms the native majority group, a scenario where there is a second-generation advantage. 3 This is the most optimistic scenario, but is also the most unrealistic, given the fact that it took about three generations for the offspring of European immigrants who arrived at the turn of the twentieth century to achieve parity with the mainstream. Second, the second generation achieves parity with, but does not outperform, the native majority group, a scenario where the second generation has experienced long-distance social mobility. This is indeed very clear and fast progress, especially by historical records. Third, the second generation outperforms native minority groups, but has yet to achieve parity with the native majority group. This is a scenario where there is a short-distance 3 Kasinitz and his colleagues used the term second-generation advantage to refer to the advantage when comparing second-generation groups to their native-born peers of the same race (e.g., Dominicans versus Puerto Ricans). They argued that it is unrealistic to expect that the second generation surpassed the native-born majority group, given the racially stratified American mainstream. We use the term second-generation advantage in a broader sense here to denote the scenario where a second-generation group outperforms the native majority group.

9 SECOND-GENERATION DECLINE OR ADVANTAGE? 9 social mobility. This suggests that the second generation is moving up, but it might take another generation or two to close the gap with the native majority group. Fourth, the second generation achieves parity with the native minority groups, a scenario where the second generation has experienced stagnation. This is a situation where there is very little progress toward upward mobility among the second generation, especially considering the relatively low level of socioeconomic attainment among the minority groups and the large racial gap in attainment in the United States. 4 Fifth, the second generation fails to achieve parity compared to the native minority groups, a scenario where there is a second-generation disadvantage. This last scenario is the most pessimistic one because it suggests either blocked mobility or downward assimilation. The second set of comparisons between the second generation and their first-generation parents informs debates on second-generation decline or progress (Gans 1992; Smith 2003; Telles and Ortiz 2008; Park and Myers 2010; Park, Myers, and Jimenez 2014). In such comparisons, we point to three possible scenarios. First, the second generation performs worse compared to their parents, a scenario where there is a second-generation decline. Second, the second generation achieves parity with their parents, a scenario where there is a second-generation stagnation. This essentially means that the assimilation process is delayed and it might take the offspring of immigrants a few more generations to achieve integration into the mainstream (Brown 2006; Bean and Stevens 2003). Third, the second generation outperforms their parents, a scenario where there is second-generation progress. These different scenarios recognize the distinction between assimilation as an outcome and assimilation as a process (Alba and Nee 2003; Marrow 2013). For example, Marrow (2013, 109) observes that assimilation is not only an outcome reflecting some convergence to a mean complete when there is no longer any discernible gap attributable to ethnicity between immigrants and their descendants and a mainstream reference group. It is also a process reflecting movement toward convergence to that mean which may occur over time or over generations, even while immigrants and their 4 Although research has identified intraclass variation among African Americans and the rise of the black middle class (e.g., Lacy 2007; Pattillo 2013), native blacks and Puerto Ricans both occupy a subordinate socio-structural position compared to native whites. At the same time, children of immigrants from minority background might draw on a minority culture of mobility to achieve upward mobility (e.g., Neckerman, Carter, and Lee 1999; Agius-Vallejo 2012; Smith 2014), which also justifies the use of native minority groups as benchmark for comparison.

10 10 INTERNATIONAL MIGRATION REVIEW descendants have not yet reached parity with, or become indistinguishable from, the members of the mainstream reference group. On the one hand, the focus on parity of outcomes might distract from the reality that assimilation is indeed occurring, despite the persistent gaps between Latino ethnic groups and their native white counterparts. On the other hand, assimilation is a process that unfolds over two or more generations through which ethnic origin becomes less and less consequential in shaping an ethnic group s life chances and outcomes. The historical records provide important lessons for our assessment of Latino progress here: European ethnic groups closed the gaps with the white Anglo-Saxon Protestant mainstream over the course of three generations (Alba and Nee 2003). Therefore, it would be unrealistic to expect Latinos to catch up in two generations, given the economic, legal, and social barriers facing them. Methodologically, we argue that extant scholarship often conflates these two sets of comparisons, leading to some confusion and slippage in both discussions of upward and downward mobility among the second generation and interpretations of the empirical evidence on these matters. Substantively, ethnic groups with human capital, legal status, and lighter skin tone (e.g., Cubans) should fare better than those with less education, lacking legal status, and facing discrimination (e.g., Mexicans or Dominicans). LATINO ASSIMILATION IN NEW DESTINATIONS One of the most important trends in settlement patterns over the last two decades is the emergence of new immigrant destinations across the country (Zu~niga and Hernandez-Leon 2005; Massey 2008; Singer, Hardwick, and Brettell 2008; Kritz, Gurak, and Lee 2011; Marrow 2011; Kritz and Gurak 2015). These new destinations encompass quite diverse contexts, ranging from rural agricultural markets and small company towns to inner-ring suburbs and rapidly developing exurbs, in all regions of the country (Marrow 2013, 108). Similarly, Singer, Hardwick, and Brettell (2008) have developed a typology of emerging, re-emerging, and pre-emerging new destinations, highlighting this very diversity. 5 For example, rural 5 Despite this diversity, the majority of Latinos in new destinations are concentrated in small towns and rural areas. In contrast, Asians specifically Indians and Chinese who are more affluent and educated are more likely to be found in more suburban new destinations. Regardless, we know that all new destinations suburban, rural, and small-town lack both institutional support and prior familiarity with immigration. As a result, immigrants in new destinations often face more challenges in their integration into American society.

11 SECOND-GENERATION DECLINE OR ADVANTAGE? 11 Latino populations are much more heavily Mexican in comparison to their Latino counterparts in traditional destinations or urban areas. Among the Latino population, the Mexican immigrant population in new destinations is more disadvantaged in socioeconomic terms not only compared to the non-mexican immigrant population, but also compared to their Mexican counterparts in traditional destinations. Put differently, Mexican Latinos are negatively selected whereas non-mexican Latinos (such as Puerto Ricans and Dominicans) are positively selected in these new destinations. The new destination context provides a unique setting for integration with both opportunities and challenges. First, Latinos in new destinations are generally more disadvantaged compared to their counterparts in the more traditional gateways (Lichter and Johnson 2009; Fischer 2010; Marrow 2013). This is the case in both rural destinations and inner-ring suburbs where Latinos have increasingly settled. Second, the gap between natives and immigrants in new destinations, at least in the small town and rural settings, tends to be more compressed, suggesting that it might be easier for the second generation to achieve parity with natives, because the benchmark for socioeconomic integration might be lower in such a context. Third, Latinos tend to face a more negative social and political reception from natives in new destinations because these new immigrant destinations often lack prior familiarity with immigration (Singer, Hardwick, and Brettell 2008; Marrow 2013). Finally, economic restructuring and job opportunities, combined with the relatively low cost of living, are among the key factors that led to the influx of Latinos in these destinations in the first place (Light 2006). And yet, there is some evidence that these opportunities are no longer as abundant in the post-recession context, posing questions about the role of new destinations in promoting upward mobility (Ellis, Wright, and Townley 2014). This would suggest that the Latino second generation will encounter a less favorable context of reception in new immigrant destinations. New immigrant destinations have fewer institutions and less infrastructure to facilitate the incorporation of immigrants and their offspring (Marrow 2011; Flippen and Parrado 2012; Marschall, Shah, and Donato 2012). Specifically, the immigrant communities in new destinations often lack the critical mass that might allow them to self-organize and provide community-based services to the immigrant population (Marrow 2013, 112). This includes general unavailability of bilingual programs in school and a lack of ready access to bilingual materials and services in local

12 12 INTERNATIONAL MIGRATION REVIEW libraries and healthcare establishments (Marrow 2011; Dondero and Muller 2012). In such a context, minority principals often play an active role in addressing the needs of immigrant and minority students (Marschall, Shah, and Donato 2012). In addition, teachers who do not share linguistic or racial/ethnic backgrounds with their students can nevertheless function in ways similar to cultural brokers [...] as a result of enhanced education, training, and professional development focused on issues of culture, language and immigration (Marschall, Shah, and Donato 2012, 147). Prior work provides no clear prediction about Latino mobility across generations in new destinations. On the one hand, some prior work would suggest that there is modest horizontal mobility among Latinos over time (Marrow 2011; Morando 2013) and the educational context can be somewhat more favorable compared to that in the old destinations (Dondero and Muller 2012). On the other hand, other studies documented the relatively lower level of educational attainment among immigrant children in new destinations (Stamps and Bohon 2006; Fischer 2010). Our study builds on this research by focusing on the Latino second generation in new destinations, a population that has begun to come of age over the last decade. DATA AND METHODS Data Sources The increasing diversity of experiences among the Latino second generation is the starting point for this project (Mitchell and Tienda 2006). In contrast to the previous decades where the majority of the Latino population was made up of Mexicans, Puerto Ricans, and Cubans, the number of both Central and South Americans has increased significantly over time. To capture this diversity, we draw on the Annual Social and Economic Supplement to the Current Population Survey (CPS ASEC), a nationally representative survey that includes a diverse Latino sample. Our empirical analyses rely on the pooled dataset from CPS ASEC 2008, 2010, and 2012, which provides the most recent data, a post-recession snapshot, and adequate sample size. The CPS ASEC is administered by the Census Bureau through both in-person and telephone interviews every month to monitor basic trends in the population. It uses a probability sample of about 60,000 occupied households from all 50 states and the District of Columbia. The survey design adopts a sampling scheme

13 SECOND-GENERATION DECLINE OR ADVANTAGE? 13 where households are included in the survey for the first four consecutive months and excluded for the next eight months, before returning for the last four consecutive months. Given this sampling design, the pooling of data from the 2008, 2010, and 2012 samples will ensure the presence of non-overlapping individuals in the pooled dataset, because each of these surveys was collected two years apart. The CPS ASEC has several strengths. First, it is currently the only data source that provides nationally representative samples of second-generation adults in the United States. Second, the pooled CPS ASEC provides for the first time adequate sample sizes for a range of Latino ethnic groups beyond Mexicans in young adulthood, in contrast to previous analyses using CPS ASEC data which have focused mostly on Mexicans (Waldinger and Feliciano 2004; Waldinger 2007; Luthra and Waldinger 2010). Third, the CPS ASEC contains adequate sample sizes, which will allow for the comparison of trends in traditional gateway versus those in new destinations. The CPS ASEC also has several limitations. First, the sample size for specific ethnic groups beyond Mexicans is rather small (Farley and Alba 2002; Smith 2003; Waldinger and Feliciano 2004; Luthra and Waldinger 2010), a limitation that this analysis will overcome by pooling CPS data from 2008 to Second, CPS ASEC data cannot address social mechanisms and processes that underlie patterns of assimilation or inequality. Third, the CPS ASEC does not contain information about parents educational and socioeconomic backgrounds, so our assessment of intergenerational mobility is indirect. We overcome this limitation by using the lagged birth cohort method (Farley and Alba 2002), a technique that allows us to directly compare outcomes of second-generation respondents with a cohort of first-generation individuals who are most likely going to be their proxy parents. With the exception of the birth cohort analyses, we will primarily focus on the second-generation experience in young adulthood. This analysis is restricted to respondents between the ages of 25 and 40 for three reasons. First, previous research mostly focuses on the new second generation in young adulthood (Kasinitz et al. 2008; Portes et al. 2005; Bean 2011). This will allow for effective comparisons of these findings to assess the results from regional surveys on the new second generation in this same age group. Second, this group of young adults has come of age under similar macrolevel economic conditions: an era of growing income inequality, expanding opportunities for higher education, liberalization of immigration policy, and official recognition of multiculturalism. Together,

14 14 INTERNATIONAL MIGRATION REVIEW these factors provide a specific context of reception for immigrants and their offspring. Focusing narrowly on this cohort of young adults eliminates confounding macrolevel factors that can contribute to differences in attainments. Dependent Variables In assessing socioeconomic assimilation among Latinos, we follow the established literature on immigrant incorporation in suggesting that social mobility often leads to integration in other spheres of social life, including proximity in both neighborhoods of residence and in the workplace, as well as inclusion in both formal and informal social networks (Alba and Nee 2003; Telles and Ortiz 2008). This analysis focuses on educational and occupational attainments among Latinos, using four dichotomous dependent variables: (1) having no high school degree, (2) having a four-year college degree or more, (3) working in a professional or managerial occupation, and (4) working in a low-level service occupation. (For the comparison between old and new destinations, we use a fifth dependent variable: working in construction). The selected occupational outcomes follow the major occupational codes from the US Census. These occupational categories capture both the low and high end of the labor market because social mobility often involves members of later generations moving away from the service sector and construction industries (i.e., working class) into administrative (i.e., lower middle class) and professional occupations (i.e., middle to upper middle class). For the birth cohort analyses, we use two continuous variables for ease of interpretation: years of education and occupational status. Our measure of occupational status is based on the occupational distribution in CPS. It is measured over a 10-point scale that maps onto the 10 major occupational codes in CPS, with managers and professionals at the top of the scale. The analyses of occupation are limited to respondents who were employed in the year prior to the survey. Analytic Methods Our analysis will focus on the largest Latino ethnic groups from the 2010 census. These include Mexicans, Cubans, Puerto Ricans, Dominicans, SGHs, CEPs, other Central Americans, and other South Americans. SGH is a composite category which includes Salvadorans, Guatemalans, and Hondurans, whereas CEP includes Colombians, Ecuadorians, and

15 SECOND-GENERATION DECLINE OR ADVANTAGE? 15 Peruvians. Central Americans and South Americans are two residual categories that include all other ethnic groups from these regions beyond SGH and CEP. Beyond practical considerations relating to sample size in the CPS data, the creation of the SGH and CEP categories are substantively justified on two grounds. First, the three ethnic groups combined in each category have quite similar socioeconomic profiles and often live in close proximity to each other, at least in New York City (Kasinitz et al. 2008). Second, the first generation among these ethnic groups faced similar contexts of exit and reception. For example, Salvadorans, Guatemalans, and Hondurans not only came from countries where political and economic conditions have been largely linked together since pre-colonial times, but they also faced significant challenges in gaining legal acceptance in the United States and reported rather high deportation rates (Menjıvar 2006; Menjıvar and Abrego 2012). That said, we recognize that there exists significant diversity even among these ethnic groups, despite their similar context of exits. In addition, CPS ASEC allows for the identification of three reference groups that include third- or higher-generation individuals (i.e., non-hispanic whites, non-hispanic blacks, and Puerto Ricans). To start, the main reference group for the multivariate analyses will be third- and higher-generation non-hispanic whites, although comparisons will also be made with third- and higher-generation non-hispanic blacks and third- and higher-generation Puerto Ricans. For brevity, we will refer to third- and higher-generation non-hispanic whites (or native whites) as the native majority group. Similarly, we will refer to third- and highergeneration non-hispanic blacks (native blacks) and third- and highergeneration Puerto Ricans as the native minority groups. (Table S1 presents the ethnic distribution in the sample). Previous theoretical frameworks provide a few straightforward predictions that guide this analysis. To begin, we compare the experience of second- and third-generation Latino ethnic groups with native whites, native blacks, and third (or higher)-generation Puerto Ricans in young adulthood. The analyses proceed in two stages. First, bivariate analyses provide statistical profiles on educational and occupational attainment for each ethnic group by ethno-racial origin, immigrant generation, and destinations. Second, multivariate logistic regression analyses explore the socioeconomic attainment of Latino ethnic groups. Because the dependent variables are dichotomous measures, we use logistic regressions with robust standard errors and report the odds ratios. Two nested models will be run

16 16 INTERNATIONAL MIGRATION REVIEW on each outcome. The first model controls only for ethno-racial origin and immigrant generation, whereas the second model further adjusts for the following covariates: age, quadratic term of age, gender, immigrant destinations, and regions. All missing values are eliminated from the analyses using listwise deletion. All analyses are adjusted for the stratified survey design using appropriate final weights provided by CPS ASEC. Operationalizing Immigrant Destinations Over the last decade, researchers have shifted the focus on traditional gateways to new immigrant destinations. Previous research has adopted various definitions of new immigrant destinations that rely on different levels of geographical units, including states, metropolitan and non-metropolitan areas, Public Use Microdata Areas (PUMAs), suburban areas, places, and counties (Waters and Jimenez 2005; Massey and Capoferro 2008; Stamps and Bohon 2006; Fischer 2010; Singer et al. 2008; Donato et al. 2008). In this paper, new immigrant destinations will be defined using county as the unit of analysis because counties provide a closest snapshot of demographic change at the local level. We chose county as the unit of analysis because this is the lowest level of geographical identifier available from the CPS ASEC sample, which allows us to classify the respondents by immigrant destination. Census data from 1990 and 2010 were then merged with the pooled CPS ASEC to classify individuals into one of these three types of counties using CPS ASEC s unique county identifier 6 (see Figure S1). For individuals without a county identifier in CPS ASEC, we created a missing category in our merging of the data. These individuals are included in the multivariate analyses to maintain our sample size, but they were labeled as missing and not classified into either new or old destinations. 7 SECOND-GENERATION ATTAINMENT IN YOUNG ADULTHOOD We begin with descriptive results on educational and occupational attainment. On education, second-generation Mexicans (16.9%) and Puerto 6 In a series of sensitivity analyses, we also explored with various thresholds for defining old versus new destinations. These analyses did not lead to substantively different results. 7 These results are not presented, but available upon request. Their inclusion and exclusion in the analyses do not provide any substantively different results.

17 SECOND-GENERATION DECLINE OR ADVANTAGE? 17 Ricans (13.6%) report the highest high school dropout rates and lowest rates of college completion. In contrast, the dropout rates among second-generation Cubans (3.6%), Central Americans (4.1%), South Americans (3.1%), and CEPs (3.8%) are similar to that of native whites (5.2%) and significantly lower than native blacks (11%). Among those with a college degree or more, Mexicans (17.4%) and Puerto Ricans (20%) reported the lowest rates, whereas the rates for Cubans (38.6%), Central Americans (42.5%), South Americans (41.1%), and CEPs (41.2%) are similar to that of native whites (38.6%) and exceed that of native blacks (20.5%). Among those who report being employed, second-generation Mexicans and Puerto Ricans are least likely to be in managerial and professional positions and more likely to be in service occupations. In contrast, second-generation Cubans, Central Americans, South Americans, and CEPs show occupational profiles more similar to that of native whites. These results are in Table S2. Figure I presents the predicted probabilities of socioeconomic attainment by ethnic groups. (The multivariate logistic regression results are in Table S3). The educational results reveal that second-generation Mexicans fare worst. As Figure I shows, they have the highest probability of being high school dropouts and the lowest probability of being graduate degree holders, in both cases registering outcomes worse than those of all three native reference groups. In contrast, second-generation Cubans, Central Americans, South Americans, and CEPs have achieved educational parity with native whites, whereas second-generation Dominicans and SGHs register better educational outcomes than native blacks and Puerto Ricans, although they are still disadvantaged compared to native whites. Are the occupational results materially different? No. Here again, second-generation Mexicans fare poorly, registering the highest probability of being in a service occupations and the lowest probability of being in professional occupations. At the other extreme, Cubans report the best outcomes, with a lower concentration in service work and a higher concentration in professional occupations, even compared to native whites. Although the other ethnic groups have yet to achieve parity with native whites, they have all surpassed native blacks and Puerto Ricans. It is remarkable that many Latino groups have achieved parity with native whites within the course of two generations. This suggests a rate of assimilation that compares well to the historical record among European groups (especially when one takes into account the many disadvantages Latinos have faced). The comparatively poor outcomes among Mexicans are likely due to their legal status and to the relatively low levels of human

18 18 INTERNATIONAL MIGRATION REVIEW Figure I. Predicted Probabilities of Socioeconomic Attainment in Young Adulthood NW3 CU2 SA2 CEP2 CA2 SGH2 DR2 PR2 MX2 MX3 PR3 NB3 NW3 CU2 SA2 CEP2 CA2 SGH2 DR2 PR2 MX2 MX3 PR3 NB3 Service Occupations Professional Occupations NW3 CU2 SA2 CEP2 CA2 SGH2 DR2 PR2 MX2 MX3 PR3 NB3 NW3 CU2 SA2 CEP2 CA2 SGH2 DR2 PR2 MX2 MX3 PR3 NB3 No High School Education College Graduate or More Notes: Results are based on the subsample of 25- to 40-year-old respondents. Horizontal lines provide the predicted probabilities for the three reference groups: native whites, native blacks, and Puerto Ricans. Predicted probabilities are based on multivariate models which also controlled for gender, age, quadratic term of age, region, and immigrant destinations (see Table 3 for full results). Abbreviations for ethnic origin are as followed: NW: native white; CU: Cuban; SA: South American; CEP: Colombian, Ecuadoran, and Peruvian; CA: Central American; SGH: Salvadoran, Guatemalan and Honduran; DR: Dominican; PR: Puerto Rican; MX: Mexican; NB: native black. The categories SGH and CEP are combined due to sample size. The residual categories of CA and SA include all the other ethnic groups from these two regions, excluding SGHs and CEPs. Source: Pooled CPS ASEC capital among the immigrant first generation, although the CPS does not contain the information needed to directly test this hypothesis. The occupational disadvantages of Mexicans may also reflect, in part, the effects of the recent recession because Mexicans are more concentrated in the service sector and construction work, industries that were severely affected by economic downturns. SOCIAL MOBILITY ACROSS IMMIGRANT GENERATIONS How do second-generation Latinos compare to their first-generation parents? It is only possible to take on this question indirectly and imperfectly. Given the cross-sectional nature of our data, first-generation Latinos in our sample are neither the actual parents of second-generation Latinos nor

19 SECOND-GENERATION DECLINE OR ADVANTAGE? 19 the actual grandparents of third-generation Latinos, making accurate assessment of intergenerational progress rather challenging. To address this limitation, we use the lagged birth cohorts approach 8 (Farley and Alba 2002; Smith 2003). Following Farley and Alba (2002), we use the lagged birth cohort method to compare a cohort of first-generation Latinos with a later cohort of second-generation Latinos who are 25 years younger. This method essentially assumes that a 25-year period approximates one immigrant generation and renders the comparison across generations more accurate. The data are arrayed so that we can meaningfully compare the average years of education and occupational status for first-generation individuals in the 1945/1965 birth cohort with those for second-generation individuals that were born 25 years later in the 1970/1990 birth cohort. We chose to focus only on these two cohorts because these two cohorts capture the first generation in middle adulthood and the second generation Latinos in young adulthood from the post-1965 period. Because Telles and Ortiz (2008) and Alba, Jimenez, and Marrow (2014) have looked at the third and higher generation with actual longitudinal data from the Mexican American Study Project, we focus here instead on the gap between the first and second generation where the CPS data are most informative. Figure II presents the average educational attainment (in years) by immigrant generation, while also showing the gap that has been closed between the first and the second generation. Among the first generation, Mexicans, along with Salvadorans, Guatemalans, and Hondurans, fare worst. Among the second generation, these two ethnic groups still report the lowest educational attainment, but the gaps between them and the other Latino groups shrank. Between the first and second generation, these two ethnic groups also reported an average gain of four years of education, the largest of all Latino groups. Although they are still at a disadvantage compared to their native peers, second-generation Mexicans and SGHS significantly outpace the parents in their educational attainment. Because their parents have only years of education on average, the children simply cannot fare worse, given compulsory education till about age 16 in the United States. For the other Latino groups, the adult children also report more education compared to their proxy parents, gaining on average about one additional year of education. On occupational attainment, Figure III shows 8 Two recent papers have also adopted the immigrant generation cohort approach (Park and Myers 2010; Park, Myers, and Jimenez 2014). We do not pursue it here because the CPS in the 1980s did not ask about nativity. Please see the Table S4 for more details.

20 20 INTERNATIONAL MIGRATION REVIEW Figure II. Educational Attainment by Ethno-Racial Origin and Generation 18 Education (in years) NW3 CU SA CEP CA SGH DR PR MX MX3 PR3 NB3 First generation Second generation Notes: Combined samples are limited to those between the ages of 25 and 65. The data are arrayed so that the results represent the average years of education for the first generation in the 1945/1965 birth cohort, along with the second generation that were born 25 years later in the 1970/1990 birth cohort, essentially assuming that the latter group is the second-generation children of the first-generation cohort. Gap bars between two data points for each ethnic group illustrate the extent of intergenerational progress by each ethnic group. Third- and higher-generation respondents are also from the 1970/1990 birth cohort. Source: Pooled CPS ASEC similarly clear progress across generation. The most disadvantaged groups report the largest intergenerational gains, especially Mexicans and SGHs. Overall, these analyses show no evidence of a second-generation decline, clear evidence of second-generation progress, and some evidence of stagnation among the third and higher generation. However, we note that the results for third- and higher generation Latinos likely underestimate the extent of assimilation among the actual third generation by averaging their outcomes with those of later-generation cohorts who are the descendants of immigrants from prior decades. For example, Alba, Jimenez, and Marrow (2014) show evidence of this happening among the fourth and higher generation using data from the Mexican American Study Project. LATINO ASSIMILATION IN OLD AND NEW DESTINATIONS How do second-generation Latinos fare in new destinations? To our knowledge, this is one of the first quantitative assessments of the Latino second generation in new destinations. We focus our discussion below on rural areas and small towns, while recognizing the diversity among new destination types. Our analysis is guided by insights from qualitative studies showing that new destinations provide a different assimilation context for Latinos compared to traditional destinations (Marrow 2011; Morando 2013). On

21 SECOND-GENERATION DECLINE OR ADVANTAGE? 21 Figure III. Occupational Attainment by Ethno-Racial Origin and Generation Occupational Status NW3 CU SA CEP CA SGH DR PR MX MX3 PR3 NB3 First generation Second generation Notes: Combined samples are limited to those between the ages of 25 and 65. The data are arrayed so that the results represent the average years of education for the first generation in the 1945/1965 birth cohort, along with the second generation that were born 25 years later in the 1970/1990 birth cohort, essentially assuming that the latter group is the second-generation children of the first-generation cohort. Gap bars between two data points for each ethnic group illustrate the extent of intergenerational progress by each ethnic group. Third- and higher-generation respondents are also from the 1970/1990 birth cohort Source: Pooled CPS ASEC the one hand, the rise of these new destinations is mostly driven by economic restructuring, with increasing demand for low-skilled workers in both the agricultural and manufacturing sectors. On the other hand, new immigrant destinations often lack the institutional resources that cater to the integration of newcomers, ranging from the availability of bilingual staff in schools and other non-profit institutions, thereby posing significant obstacles for integration. With regard to social mobility, inequality between the native-born and immigrant populations is more compressed in new immigrant destinations. In such a context, even short-distance social mobility can be consequential and bring the second generation closer to parity with the mainstream. For this analysis, we will focus primarily on Mexicans because of their significant presence in the new destinations, while also grouping all the other Latino groups into one composite category non-mexican Latinos due to sample size consideration. 9 The sample is limited to first- and 9 The number of second-generation Latino respondents in new destinations is as followed: Mexicans (N = 94), Puerto Ricans (N = 82), and all other Latinos (N = 58). We recognize the small sample size for the non-mexican ethnic groups, so our analyses focused on second-generation Mexicans (N = 94) versus non-mexicans (N = 140).

22 22 INTERNATIONAL MIGRATION REVIEW second-generation Latino respondents between the ages of 25 and 40, in comparison to the third- and higher-generation respondents from the reference groups. We begin with the descriptive statistics by ethnic group and destination. The overall pattern is clear: The outcomes for all groups are slightly worse in new destinations compared to those in traditional gateways, with those in new destinations reporting lower levels of education, higher concentrations in construction work, and lower concentrations in managerial and professional occupations. Among second-generation Mexicans, 17.3 percent are college graduates and 21.9 percent are professionals in traditional gateways, whereas these proportions are only 13.8 percent and 16 percent in new destinations. In contrast, these rates are much higher among both native whites and native blacks. These results are in Table S5. Do second-generation Mexicans and non-mexicans fare better in traditional gateways than new destinations? We test the relative difference in attainment between old and new immigrant destinations by ethnic group by introducing a series of interaction terms for destinations and ethnic origin in a multivariate regression framework. These results are in Table S6. To facilitate interpretation and comparisons of these results, we estimated the marginal probabilities based on the interaction effects from the multivariate models in Table S6 for a 25-year-old male respondent in the sample, with old destinations as the reference category. In substantive terms, the marginal probabilities indicate the probability for a specific outcome to occur for a 25-year-old male respondent in new destination compared to a 25-year-old male respondent in a traditional gateway from the same ethno-racial group. These results are reported in Table 1. Controlling for observable covariates, native whites, native blacks, and first-generation Mexicans are all negatively selected on the four indicators in new destinations compared to traditional gateways. Specifically, those in new destinations are significantly more likely to have no high school education and work in construction compared to their counterparts from the same ethno-racial and generational group in old destinations. Similarly, those in new destinations are also significantly less likely to be college graduates or professionals compared to their counterparts from the same ethno-racial and generational group in traditional gateways. These results generally confirm findings from the descriptive statistics in Table S5 and are consistent with research documenting the role of economic opportunities in low-wage work in attracting first-generation Latinos to new destinations. Among both second-generation Mexicans

23 SECOND-GENERATION DECLINE OR ADVANTAGE? 23 TABLE 1 MARGINAL PROBABILITIES OF SOCIOECONOMIC ATTAINMENT IN NEW VERSUS OLD DESTINATIONS No High School College Graduate Margin SD Margin SD White, third and higher generation 0.018*** *** Black, third and higher generation 0.024* ** Mexican, first generation * Mexican, second generation Mexican, third and higher generation Non-Mexican, first generation 0.121*** Non-Mexican, second generation Puerto Rican, third and higher generation Construction Work Managerial/Professional Margin SD Margin SD White, third and higher generation 0.019** *** Black, third and higher generation 0.034** *** Mexican, first generation 0.130*** *** Mexican, second generation Mexican, third and higher generation Non-Mexican, first generation 0.096* Non-Mexican, second generation Puerto Rican, third and higher generation Notes: Results are based on the subsample of 25-to-40-year-old respondents. Marginal probabilities are based on the interaction effects from the multivariate models which also controlled for gender, age, and immigrant destination. The reference group is old destinations. Marginal probabilities were estimated for a 25-year-old male respondent. ***p < 0.001, **p < 0.01, *p < 0.05, p < Source: Pooled CPS ASEC ( ). and non-mexicans, there are no statistically significant differences by immigrant destinations. That said, the presence of the second generation in new destinations is still very small, so a full assessment will have to await future research. More broadly, these results reflect the occupational structures of new destinations, which lean more heavily toward the farming business, the meat-packing industry, and construction work, given these areas more rural and small-town context. In contrast, both professional and service jobs are more prevalent in urban metropolitan centers. ASSIMILATION INTO WHAT AND PROGRESS COMPARED TO WHOM? The questions of assimilation into what and progress compared to whom remain central to our assessment of the contentious debate on second-generation outcomes. Given the racially stratified nature of American society, these questions are non-trivial ones. Table 2 attempts to bring

24 24 INTERNATIONAL MIGRATION REVIEW TABLE 2 LATINO SECOND-GENERATION ATTAINMENT COMPARED TO THREE REFERENCE GROUPS High School or Less College Degree or More Education NW3+ NB3+ PR3+ NW3+ NB3+ PR3+ Mexican Puerto Rican Dominican SGH Central Americans CEP South Americans Cuban Low-Level Service Managerial and Professional Occupation NW3+ NB3+ PR3+ NW3+ NB3+ PR3+ Mexican Puerto Rican Dominican SGH Central Americans CEP South Americans Cuban Notes: The three main reference groups are native whites, native blacks, and Puerto Rican, third- and higher generation. Comparisons of relative advantage or disadvantage in a particular outcome are based on the odds ratios for the relevant second-generation group from the multivariate analyses with the corresponding reference group, adjusting for observable covariates in the full models. For negative outcomes (e.g., having less than a high school education), odds ratios that are larger than 1 will be noted as a disadvantage. In other words, if second-generation Mexicans are more likely than native whites to have no high school degree, this is an actual disadvantage. For positive outcomes (e.g., being in a professional occupation), odds ratios that are larger than 1 will be noted as an advantage. In other words, if second-generation Cubans are more likely than native blacks to be in a professional or managerial position, this is an actual advantage. All of these results are summarized here for the Latino second generation only. Specifically: + denotes a significant advantage of a second-generation group over a native group; denotes a significant disadvantage of a second-generation group relative to a native group; 0 denotes no significant differences between a second-generation group and a native group. some clarity to this debate by assessing second-generation attainment in relations to the three reference groups: native whites, native blacks, and Puerto Ricans. These three reference groups present different anchoring points, with parity with native whites on the one end and parity with native blacks on the other end of the same continuum of progress. Our work builds directly on the work of Park, Myers, and Jimenez (2014) and their call for more contextualized benchmark to assess the second-generation progress against a constantly changing American mainstream. In comparison to native whites, Table 2 shows that second-generation Cubans, CEPs, Central Americans, and South Americans have

25 SECOND-GENERATION DECLINE OR ADVANTAGE? 25 achieved parity in educational and occupational attainment, whereas second-generation Mexicans, Puerto Ricans, Dominicans, and SGHs continue to lag behind. In comparison to native blacks, Mexicans and Puerto Ricans reported lower levels of educational attainment, but higher levels of occupational attainment, suggesting that these groups have been either somewhat more successful than native blacks to translate their educational credential into professional gains in the labor market or more concentrated in occupations that do not require educational credentials, such as construction and service work. The ongoing structural discrimination and prejudice toward native blacks no doubt play a role here, but these results also underscore the relative progress among second-generation Mexicans. 10 The remaining Latino ethnic groups have either achieved parity or reported an advantage over native blacks on these measures of attainment. In comparison to third- and higher-generation Puerto Ricans, second-generation Mexicans fare slightly worse educationally, while reporting similar occupational profiles. What this exercise underscores is how different conclusions about second-generation progress and assimilation can depend on the reference population. What these results also show is the unmistakable progress among some of the Latino second generation, even when we use native whites as the yardstick of assessment. We now revisit the two questions at the core of our analyses. First, are second-generation Latinos at a disadvantage or an advantage compared to their native peers? Second, are second-generation Latinos experiencing upward mobility or are they at risk of downward mobility compared to their parents? Table 3 brings clarity to these questions by summarizing the paper s main results. First, we compare outcomes among the Latino second generation with their native minority and native majority peers. Then, we compare them with their proxy parents. On the first question, second-generation Mexicans and Puerto Ricans are at a relative disadvantage, even when we use native minority groups as yardsticks of comparison. Dominicans are next in the order, reporting outcomes that are on par with the native minority groups. Above them, Salvadorans, Guatemalans, and Hondurans have outperformed the native 10 However, first-generation Mexicans also exhibit higher occupational incorporation than native blacks, despite their relatively lower educational attainment and English proficiency. This would suggest that first-generation Mexicans are also more likely than native blacks to translate their human capital into labor market success.

26 26 INTERNATIONAL MIGRATION REVIEW TABLE 3 CONCEPTUAL SCHEMAS OF MOBILITY PROCESSES AMONG SECOND-GENERATION LATINOS Mobility Process Reference Groups Latino Ethnic Groups Disadvantage Worse than native minority MX, PR Stagnation Parity with native minority DR Short-distance mobility Better than native minority, SGH but worse than native majority Long-distance mobility Parity with native majority CU, SA, CEP, CA Advantage Better than native majority None Mobility Process Reference Groups Latino Ethnic Groups Downward or decline Worse than own parents None Horizontal or stagnation Parity with own parents MX3+, PR3+ Upward or progress Better than own parents CU, SA, CEP, CA, DR, SGH, MX Notes: Abbreviations denote second-generation groups. The exceptions are MX3+ and PR3+, which denote thirdand higher-generation Mexicans and Puerto Ricans. minority groups, but have yet to achieve parity with the native majority group, so their mobility pattern is best described as short-distance mobility. In contrast, Cubans, Colombians, Ecuadorians, Peruvians (CEP), Central Americans, and South Americans have achieved parity with the native majority group, essentially closing the gaps within the course of two generations, so their mobility pattern is best described as long-distance mobility. At the same time, there is no evidence of a second-generation advantage compared to the native majority group at this time. These results not only do not invalidate, but also lend support for findings from the major study of the second generation in New York City, which documented a second-generation advantage over native groups of the same race (Kasinitz et al. 2008). In other words, Kasinitz and his colleagues were right in asserting that it is more appropriate to compare outcomes among the second generation to a native group of the same race because the socioeconomic attainment of the native majority group (i.e., native white) is an unreasonably high benchmark to assess secondgeneration progress, given the racially stratified nature of American society. On the second question, there is no evidence of a second-generation decline and downward mobility, as none of the Latino groups report worse socioeconomic outcomes compared to their proxy first-generation parents. Quite the contrary, the evidence points to clear intergenerational progress and upward mobility among the Latino second generation compared to their parents. That said, third- and higher-generation Puerto Ricans and Mexicans show some signs of stagnation, reporting very little progress beyond the second generation.

27 SECOND-GENERATION DECLINE OR ADVANTAGE? 27 DISCUSSION AND CONCLUSION Given the sheer demographic presence of Latinos, assessing their assimilation provides key insights into the continuing transformation of American society in the coming decades. Since the onset of the global economic recession in December 2007, there has been increasing evidence that Latinos and African Americans were hit the hardest, given their precarious situation and their tenuous position in the labor market. As a result, understanding social mobility among Latinos has become even more pressing in an era of growing inequality, economic recession, stagnation of wages, and increasing financial insecurity (Grusky, Western, and Wimer 2011). By contrast, previous studies that focus on different time periods would likely capture different Latino populations who might have come of age under vastly different economic conditions and structures of opportunities (Smith 2003; Telles and Ortiz 2008; Alba, Jimenez and Marrow 2014; Park, Myers, and Jimenez 2014), making this analysis of post-recession trends on attainment among Latinos all the more relevant. Furthermore, the rise of new immigrant destinations has raised concerns over the mobility trajectory of the second generation and the changing context of assimilation (Marrow 2013; Morando 2013). In light of these trends, we provide one of the first post-recession snapshots of the Latino experience in both old and new immigrant destinations. Our assessment of Latino socioeconomic assimilation is an optimistic one, showing clear progress among the Latino second generation. These results also highlight how our choice of benchmark to assess second-generation progress really matters. The literature on assimilation and social mobility routinely conflates comparisons of second-generation outcomes with those of their parents and with their native peer groups. As we have clearly show, these two sets of comparisons not only lead to very different conclusions about second-generation progress, but also are driven by different mechanisms and assume very different benchmarks for evaluating second-generation progress. On the one hand, the debate on second-generation disadvantage and advantage, along with the possibilities of short-distance mobility and long-distance mobility, is often based on the assumption of a homogenous American mainstream into which the second generation can be incorporated. However, the choice of native reference group to benchmark second-generation outcomes is equally important because the native majority group (i.e., non-hispanic whites) is at a signif-

28 28 INTERNATIONAL MIGRATION REVIEW icant advantage over the native minority groups (i.e., non-hispanic blacks and Puerto Ricans). This choice is not just a mere methodological concern, but also reflects the changing meaning of assimilation in a diverse American society where assimilation into the American mainstream no longer means achieving parity with native whites, especially in light of the growth of the black and Latino middle class and the black elite (Lacy 2007; Agius-Vallejo 2012; Pattillo 2013). On the other hand, the debate on second-generation decline or progress is often based on the assumption of intergenerational progress over time, a scenario in which the children of immigrants are expected to achieve more socioeconomic mobility and become more integrated into the American mainstream compared to their foreign-born parents. This is an assumption that is not only deeply rooted in the theory of assimilation that grew out of the experience of European immigrants who arrived a century ago, but also might be untenable given declining economic opportunities in the aftermath of the recession and in the era of growing income inequality. As a result, this also raises the possibility of stagnation or even decline across immigrant generations, a possibility that we find theoretically intriguing, but with no empirical support. Furthermore, the human capital profile of post-1965 immigrants differs dramatically from that of European immigrants who arrived a century ago, sometimes rendering intergenerational progress unrealistic. As Gans (2007, 152) has observed: the immigration researchers who first studied [the European immigrants] took their upward mobility for granted and focused on their assimilation. Although socioeconomic assimilation is an important part of the overall assimilation process, social mobility and assimilation are conceptually distinct. For example, if someone migrates to the United States from South America with an advanced degree, is it realistic for us to expect her second-generation children to exceed her level of education? In this example, the possibility of a second-generation decline is a real one. We also note the diversity in mobility outcomes among Latinos, ranging from a second-generation disadvantage and stagnation among some groups to short- and long-distance mobility among others. In other words, the Latino experience does not fit neatly into either the assimilation or racialization framework. Even among Mexicans, one of the most racialized Latino groups, there is clear heterogeneity both across generations and across regional contexts, with legal status being an important dividing line for their eventual integration (Telles and Ortiz 2008; Alba, Jimenez, and Marrow 2014; and Park, Myers, and Jimenez 2014).

29 SECOND-GENERATION DECLINE OR ADVANTAGE? 29 These results are consistent with findings from Telles and Ortiz (2008, 17) which also showed that [s]ome Mexican Americans have done well and most do better now than their parents did when they were young in the 1960s. Those are signs of partial assimilation. However, they continue to lag well behind their Anglo counterparts, which also reflects a process of racialization. We extend this line of argument to the remaining Latino ethnic groups in our sample by documenting similar heterogeneity in their integration into American society, while also showing clear clustering of groups into the specific mobility processes outlined in Table 3. Although our paper primarily focuses on the second generation, we also find some evidence of stagnation or lack of progress among third- and higher-generation Puerto Ricans (and to a lesser extent the third- and higher-generation Mexicans). Our findings confirm prior evidence on third- and higher-generation Mexicans from Telles and Ortiz (2008, 16), which also showed that economic integration, the most desirable aspect of assimilation, stalls after the second generation while cultural, social and political assimilation occurs slowly but constantly over generations-since immigration. One potential explanation for this is the composition of third-generation individuals who self-identify as Mexican or Puerto Rican (Alba and Islam 2009). Put differently, upwardly mobile thirdgeneration Latinos with lighter skin tone might no longer identify ethnically, choosing instead to self-identify as white. The selective attrition would suggest caution in our interpretation of a third-generation decline (Duncan and Trejo 2011). Another potential explanation is the unbridled drive and determination typically seen among first-generation immigrants and their second-generation offspring, which Smith (2006) and Louie (2012) refer to as the immigrant bargain, might be less of a driving force behind third-generation attainment. Finally, new immigrant destinations provide a specific context for integration where the inequality in outcomes between immigrants and natives is relatively more compressed (Marrow 2013), suggesting that it might be easier for the second generation to close this gap if they have access to higher education, employment trainings, bilingual programs, and other social resources. Although we expect mobility to be more modest in new immigrant destinations, we find no difference in outcomes among second-generation Mexicans across destinations, an important finding in light of the changing geography of immigration. This also confirms prior qualitative research that showed modest horizontal mobility among Latinos in

30 30 INTERNATIONAL MIGRATION REVIEW new destinations, while suggesting the need for more support in facilitating their incorporation (Marrow 2011). This set of analyses also has the following limitations. First, this article cannot directly compare second-generation Latinos to their own first-generation parents, as CPS ASEC data do not allow for such an analysis. Second, this article cannot account for social class backgrounds of the respondents or human capital differences across groups among the first generation. If such data are available, it is entirely possible that second-generation Latinos with social class backgrounds similar to their native counterparts might report even better outcomes compared to those of the native groups. This would lend even stronger support to the conclusions we draw here. Third, this analysis focuses on the differences in outcomes across ethno-racial groups and immigrant generations, but cannot speak of the social processes and the mechanisms that underlie the patterns documented. Moreover, we cannot speak of the role of family values, cultural beliefs, and aspirations in shaping these divergent sets of outcomes (Kao and Thompson 2003; Kasinitz et al. 2008). Fourth, this analysis cannot take into account legal status, an increasingly important dimension of stratification over the last decade (Bean 2011; Gonzales 2011). In fact, the lack of legal status is likely to be a significant factor among Mexicans, who comprise the majority of the 11.2 million undocumented individuals, especially in light of research that shows a clear negative impact of parents and grandparents lack of legal status on educational attainment. Finally, these findings also highlight the important role of skin color and racial phenotype in shaping the stratification process (Telles 2014). Because Latinos are neither black nor white, how much skin color accounts for the intra-latino variation in outcomes across Latino ethnic groups remains an important open question for future research. These analyses yield an optimistic assessment of second-generation progress in this post-recession evaluation, while noting the disadvantages and vulnerability among Mexicans and Puerto Ricans. Overall, Latinos are increasingly being integrated into the American society. The Latino second generation made striking gains on education and moved away from the low-paying occupations prevalent among the Latino first generation. Given the importance of education as the linchpin of slow assimilation among third- and fourth-generation Mexicans (Telles and Ortiz 2008, 274), these findings give us reason to be optimistic. At the same time, some groups have yet to achieve full parity with native whites and to

31 SECOND-GENERATION DECLINE OR ADVANTAGE? 31 secure a place in the highest echelons of the American mainstream. Although the future of Latinos as a group remains uncertain, we have reasons to be optimistic because they have weathered the recent economic downturn well and are poised to take advantage of the new opportunities that lie ahead. REFERENCES Agius-Vallejo, J Barrios to Burbs: The Making of the Mexican-American Middle Class. Palo Alto: Stanford University Press. Alba, R. D., T. R. Jimenez, and H. B. Marrow 2014 Mexican Americans as a Paradigm for Contemporary Intra-Group Heterogeneity. Ethnic & Racial Studies 37(3): , and T. Islam 2009 The Case of the Disappearing Mexican Americans: An Ethnic-Identity Mystery. Population Research & Policy Review 28(2): , P. Kasinitz, and M. C. Waters 2011 The Kids Are (Mostly) Alright: Second-Generation Assimilation. Social Forces 89 (3): , and V. Nee 2003 Remaking the American Mainstream: Assimilation & Contemporary Immigration. Cambridge: Harvard University Press. Bean, F. D., M. A. Leach, S. K. Brown, J. D. Bachmeier, and J. R. Hipp The Educational Legacy of Unauthorized Migration: Comparisons Across U.S.- Immigrant Groups in How Parents Status Affects Their Offspring. International Migration Review 45(2): Brown, S.K Delayed Spatial Assimilation: Multi-Generational Incorporation of the Mexican- Origin Population in Los Angeles. City & Community 6: Bean, F.D., and G. Stevens 2003 America s Newcomers and the Dynamics of Diversity. New York: Russell Sage Foundation. Donato, K. M., C. Tolbert, A. Nucci, and Y. Kawano 2008 Changing Faces, Changing Places: The Emergence of New Non-Metropolitan Immigrant Gateways. In New Faces in New Places: The Changing Geography of American Immigration, edited by D. S. Massey, New York: Russell Sage Foundation. Dondero, M., and C. Muller 2012 School Stratification in New & Established Latino Destinations. Social Forces 91 (2): Duncan, B., and S. Trejo 2011 Tracking Intergenerational Progress for Immigrant Groups: The Problem of Ethnic Attrition. American Economic Review 101(3): Ellis, M., R. Wright, and M. Townley 2014 The Allure of New Immigrant Destinations & the Great Recession in the United States. International Migration Review 48(1):3 33.

32 32 INTERNATIONAL MIGRATION REVIEW Farley, R., and R. Alba 2002 The New Second Generation in the United States. International Migration Review 36(3): Fischer, M. J Immigrant Educational Outcomes in New Destinations: An Exploration of High School Attrition. Social Science Research 39(4): Flippen, C., and E. A. Parrado 2012 Forging Hispanic Communities in New Destinations: A Case Study of Durham, NC. City & Community 1:1 30. Gans, H. J Acculturation, Assimilation & Mobility. Ethnic & Racial Studies 30(1): Second-Generation Decline: Scenarios for the Economic & Ethnic Futures of the Post-1965 American Immigrants. Ethnic & Racial Studies 15(2): Gonzales, R. G Learning to Be Illegal: Undocumented Youth & Shifting Legal Contexts in the Transition to Adulthood. American Sociological Review 76(4): Greenman, E., and M. Hall 2013 Legal Status & Educational Transitions for Mexican & Central American Immigrant Youth. Social Forces 91(4): Grusky, D. B., B. Western, and C. Wimer 2011 The Great Recession. New York: Russell Sage Foundation. Haller, W. J., A. Portes, and S. Lynch 2011 Dreams Fulfilled, Dreams Shattered: Determinants of Segmented Assimilation in the Immigrant Second Generation. Social Forces 89(3): Jimenez, T. R Replenished Ethnicity: Mexican Americans, Immigration & Identity. Berkeley: University of California Press. Kao, G., and J. Thompson 2003 Racial & Ethnic Stratification in Educational Achievement & Attainment. Annual Review of Sociology 29: Kasinitz, P., J. H. Mollenkopf, M. C. Waters, and J. Holdaway 2008 Inheriting the City: The Children of Immigrants Come of Age. New York: Russell Sage Foundation. Kritz, M. M., and D. T. Gurak 2015 U.S. Immigrants in Dispersed and Traditional Settlements: National Origin Heterogeneity. International Migration Review 49(1): ,, and M. A. Lee 2011 Will They Stay? Foreign-Born Out-Migration from New U.S. Destinations. Population Research & Policy Review 30: Lacy, K Blue-Chip Black: Race, Class, & Status in the New Black Middle Class. Berkeley: University of California Press. Lee, J., and F. D. Bean 2010 The Diversity Paradox: Immigration & The Color Line in 21st Century America. New York: Russell Sage Foundation Press. Lichter, D. T., and K. M. Johnson 2009 Immigrant Gateways & Hispanic Migration to New Destinations. International Migration Review 43(3):

33 SECOND-GENERATION DECLINE OR ADVANTAGE? 33 Light, I Deflecting Migration: Networks, Markets and Regulation in Los Angeles. New York: Russell Sage Foundation. Lopez, M. H., and G. Velasco 2011 Childhood Poverty among Hispanics Sets Record, Leads Nation. Washington: Pew Hispanic Center. Louie, V Keeping the Immigrant Bargain: The Costs & Rewards of Success in America. New York: Russell Sage Foundation. Luthra, R. R., and R. Waldinger 2010 Into the Mainstream? Labor Market Outcomes of Mexican Origin Workers. International Migration Review 44(4): Marrow, H. B Assimilation in New Destinations. Dædalus: The Journal of the American Academy of Arts & Sciences: 142(3): New Destination Dreaming: Immigration, Race & Legal Status in the Rural American South. Stanford: Stanford University Press. Marschall, M., P. Shah, and K. Donato 2012 Parent Involvement Policy in Established & New Immigrant Destinations. Social Science Quarterly 93(1): Massey, D. S New Faces in New Places: The Changing Geography of American Immigration. New York: Russell Sage Foundation., and C. Capoferro 2008 The Geographic Diversification of U.S. Immigration. In New Faces in New Places: The Changing Geography of American Immigration, edited by D. S. Massey, New York: Russell Sage Foundation., J. Durand, and N. J. Malone 2002 Beyond Smoke & Mirrors: Mexican Immigration in an Era of Economic Integration. New York: Russell Sage Foundation. McKernan, S., C. Ratcliffe, E. Steuerle, and S. Zhang 2013 Less Than Equal: Racial Disparities in Wealth Accumulation. Washington: The Urban Institute. Menjıvar, C., and L. J. Abrego 2012 Legal Violence: Immigration Law & the Lives of Central American Immigrants. American Journal of Sociology 117(5): Liminal Legality: Salvadoran & Guatemalan Immigrants Lives in the United States. American Journal of Sociology 111(4): Molly, D., and C. Muller 2012 School Stratification in New and Established Latino Destinations. Social Forces 91(2): Morando, S. J Paths to Mobility: The Mexican Second Generation at Work in a New Destination. The Sociological Quarterly 54(3): Neckerman, K. M., P. Carter, and J. Lee 1999 Segmented Assimilation & Minority Cultures of Mobility. Ethnic & Racial Studies 22(6):

34 34 INTERNATIONAL MIGRATION REVIEW Nicholls, W. J The DREAMers: How the Undocumented Youth Movement Transformed the Immigrant Rights Debate. Stanford: Stanford University Press. Park, J., D. Myers, and T. R. Jimenez 2014 Intergenerational Advancement of the Mexican-origin Population in California & Texas Relative to a Changing U.S. Mainstream. International Migration Review 48 (2): , and D. Myers 2010 Intergenerational Mobility in the Post-1965 Immigration Era: Estimates by an Immigrant Generation Cohort Method. Demography 47: Passel, J. S., and D V. Cohn 2011 Unauthorized Immigrant Population: National & State Trends, Washington: Pew Hispanic Center. Pattillo, M Black Picket Fences: Privilege & Peril Among the Black Middle Class, 2nd ed. Chicago: University of Chicago Press. Perlmann, J Italians Then, Mexicans Now: Immigrant Origins & Second-Generation Progress, 1890 to New York: Russell Sage Foundation. Portes, A., P. Fernandez-Kelly, and W. Haller 2005 Segmented Assimilation on the Ground: The New Second Generation in Early Adulthood. Ethnic and Racial Studies 28(6): Portes, A., and R. G. Rumbaut 2001 Legacies: The Story of the Immigrant Second Generation. Berkeley; New York: University of California Press., and M. Zhou 1993 The New Second Generation: Segmented Assimilation & its Variants. The Annals of the American Academy of Political & Social Science 530(1): Roth, W. D Race Migrations: Latinos & the Cultural Transformation of Race. Stanford: Stanford University Press. Rendon, M. G Caught Up: How Urban Violence & Peer Ties Contribute to High School Non- Completion. Social Problems 61(1): Singer, A., S. W. Hardwick, and C. Brettell 2008 Twenty-First Century Gateways: Immigrant Incorporation in Suburban America. Washington: Brookings Institution Press. Smith, R. C Black Mexicans, Conjunctural Ethnicity, & Operating Identities: Long-Term Ethnographic Analysis. American Sociological Review 79(3): Mexican New York: Transnational Lives of New Immigrants. Berkeley: University of California Press. Smith, J. P Assimilation Across the Latino Generations. American Economic Review 93 (2): Stamps, K., and S. Bohon 2006 Educational Attainment in New & Established Latino Metropolitan Destinations. Social Science Quarterly 87:

35 SECOND-GENERATION DECLINE OR ADVANTAGE? 35 Telles, E Pigmentocracies: Ethnicity, Race, and Color in Latin America. Chapel Hill: University of North Carolina Press. Telles, E., and V. Ortiz Generations of Exclusion: Mexican Americans, Assimilation & Race. New York: Russell Sage Foundation. Tienda, M., and F. Mitchell 2006 Multiple Origins, Uncertain Destinies: Hispanics & the American Future. Washington: National Academy Press. US Department of Homeland Security Yearbook of Immigration Statistics: Washington: US Department of Homeland Security, Office of Immigration Statistics. US Census Bureau Profile America: Facts for Features: Hispanic Heritage Month Washington: US Census Bureau. Vaquera, E., E. Aranda, and R. Gonzales 2014 Patterns of Incorporation of Latinos in Old & New Destinations: From Invisible to Hypervisible. American Behavioral Scientist 58(14): Waldinger, R Did Manufacturing Matter? The Experience of Yesterday s Second Generation: A Reassessment. International Migration Review 41(1):3 39., and C. Feliciano 2004 Will the New Second Generation Experience Downward Assimilation? Segmented Assimilation Re-Assessed. Ethnic & Racial Studies 27(3): Waters, M. C Defining Difference: The Role of Immigrant Generation & Race in American & British Immigration Studies. Ethnic & Racial Studies 37(1):10 26., and T. R. Jimenez 2005 Assessing Immigrant Assimilation: New Empirical & Theoretical Challenges. Annual Review of Sociology 31: Zu~niga, V., and R. Hernandez-Leon 2005 New Destinations: Mexican Immigration in the United States. New York: Russell Sage Foundation. SUPPORTING INFORMATION Additional supporting information may be found in the online version of this article at the publisher s web site: Figure S1. Old and New Immigrant Destinations by County, Table S1 Ethnic Groups by Immigrant Generation. Table S2 Socioeconomic Attainment by Ethnic Group. Table S3 Logistic Regressions on Educational Attainment in Young Adulthood.

36 36 INTERNATIONAL MIGRATION REVIEW Table S4 Educational Attainment among First-Generation Latinos from Three CPS Samples. Table S5 Socioeconomic Assimilation in Old and New Immigrant Destinations. Table S6 Logistic Regressions on Socioeconomic Attainment in New and Old Destinations.

37 Figure S1: Old and New Immigrant Destinations by County, Notes: Following Donato et al. (2010), we use county-level census data to classify counties into traditional destinations, new destinations, or other destinations, spanning two decades. Specifically, we draw on the Decennial Censuses (1990 and 2000), as well as data from the American Community Survey ( ) to benchmark growth and to classify counties because these two decades witnessed the most dramatic growth in new immigrant destinations (Lichter and Johnson 2009). Traditional destinations are defined as counties that meet the following criteria: (1) the Hispanic population at the county level is at least 5%, and (2) the Hispanic population in 1990 was at least 1,000. New destinations are defined in two stages based on Hispanic population size and growth rate between and For the decade, new destinations are counties where: (1) the Hispanic population at the county level was below 5% in 1990, (2) the growth rate of the Hispanic population was at least 90% between 1990 and 2000, and (3) the Hispanic population in the county was at least 1,000 in the year For the decade, new destinations are counties in which: (1) the Hispanic population at the county level was below 5% in 2000, (2) the growth rate of the Hispanic population was at least 75% between 2000 and 2010, and (3) the Hispanic population in the county was at least 1,000 in Other destinations include the remaining counties in the country that started out with a below-average foreign-born population in 1990 and have experienced either modest or negative growth between 1990 and This is generally consistent with the approaches of previous research, with two improvements: (1) we rely on county-level data instead of data derived from states or metropolitan areas, and (2) we use Hispanic composition within the county instead of just immigrant composition. Specifically, the classification schema was based on 3,134 counties in the U.S. with complete demographic data from both 1990 and In total, 8.07% of the counties were classified as traditional destinations, 22.27% were classified as new destinations and 69.66% were classified as other destinations.