Arthur Sweetman
Khuong Truong
Catching Up? Country Studies on Intergenerational Mobility and Children of Immigrants
Chapter 7. The United States and Canada: Intergenerational social mobility among immigrants and their native-born children
Abstract
This chapter reviews the academic literature on new immigrants’ intergenerational educational and labour market integration in the United States and Canada, and presents new findings. It begins with a discussion of American and Canadian immigration history, and then addresses the intergenerational transmission of educational outcomes. Particular attention is paid to students from disadvantaged backgrounds, especially minority ethnic-group students. The discussion points to language deficiencies as a major drawback, and outlines possible reasons for the relatively slow integration of the Hispanic community into higher levels of education in the United States. The chapter then turns to labour market outcomes in both countries, examining participation rates and earnings gaps for adult immigrants, immigrants who arrived as children, the children of immigrants and, as a comparison group, children with two native-born parents.
Main findings
The following key observations are based on a review of the academic literature and data from the OECD’s Programme for the International Assessment of Adult Competencies (PIAAC). For the purposes of this study, unless otherwise indicated, the entire population of permanent residents is divided into four mutually exclusive categories. “Immigrant” refers to individuals who permanently migrate to either the United States or Canada as adults (or as older youth). Regardless of their current age, “immigrant children”, refers to persons who immigrated at a young age (the age threshold varies across studies, but commonly before age 11 or so) and “children of immigrants” refers to those native-born children who have at least one immigrant parent. A distinct group, normally used as a comparison, comprises the children of parents who are both native born, i.e., the “children of the native born” or “children of non-immigrants”. Unless otherwise specified, all measurements of educational and labour market outcomes are made when these “children” are adults. That is, for example, a 60-year old who immigrated at age four is an immigrant child.
The shift from traditional immigration source countries to a larger set of more global source countries that occurred beginning in the 1960s for both the United States and Canada has had implications for the native-born children of those immigrants. For example, children of this “new” immigration from Asia have a remarkably high university attendance rate that is relatively insensitive to traditional predictors of postsecondary attendance, such as parents’ education, family income and high school marks.
One important phenomenon of the more recent immigration in the United States is the high percentage of persons who emigrated from Mexico and to a lesser extent from Central America. In 2011, Mexican immigration accounted for almost 30% of the total immigrant stock in the United States. In contrast, immigration to Canada is more diversified.
For the United States and even more so for Canada, the correlation between children’s educational attainment and that of their parents is much weaker for immigrant parents than for native-born parents. The educational trajectory of children of immigrant parents is less determined by their parents’ level of education.
In the United States, there is a substantial increase in average educational attainment between immigrants from Mexico and their children, but from a very low base. The average educational attainment of the children of Mexican immigrants is around the end of high school, which is not very far from the education level associated with the compulsory schooling age.
On average, in both the United States and Canada the educational attainment of the children of immigrants is equal to or exceeds that of same-aged children of native-born parents. Female children of immigrants in the United States have years of schooling comparable to that of female children of native-born parents, but for males in the United States and for both genders in Canada the children of immigrants accumulate on average between about one-half and three-quarters of a year of additional schooling.
Most remarkable is that among all ethnic groups, with the exception of those whose parents arrived from Central or South America or the Caribbean, children of immigrants have a higher probability of attending university than the children of the native born. Especially the ethnic groups most linked with postsecondary attendance (such as the children of Chinese immigrants) seem to enrol in postsecondary education even when they have relatively low marks in high school. Furthermore, postsecondary attendance is relatively insensitive to family income for the children of immigrants in many ethnic groups. Across groups, postsecondary access appears not so much the major issue as is postsecondary completion.
Turning to earnings, in the United States the children of immigrants have earnings that are essentially indistinguishable from those of the children of natives, with or without controlling for observable characteristics.
In Canada, the children of immigrants have appreciably higher earnings than the children of natives. This is mostly due to the higher educational attainment of immigrants’ children and their location in highly urbanised areas with earning premiums.
Introduction
Immigrant recipient societies need be concerned about the success of the children of new arrivals. Of particular relevance are the accumulation of skills by those children that are useful in the labour market – especially education and domestic language skills – and the children’s labour market outcomes as adults. This chapter documents aspects of the current understanding of these intergenerational processes among immigrants1 and their children in the United States and Canada, with the focus on disadvantaged groups. Recent US-Canada comparisons of immigrant intergenerational integration are by Picot and Hou (2011a, b), and a survey covering a broader set of issues related to the children of immigrants is by Sweetman and van Ours (2015).
The United States and Canada took different paths with respect to immigration after the Second World War, and again in the 1960s. This has led not only to a much higher immigration rate, but also to a different demographic and educational composition for new immigrants. Additionally, as geography has it, Canada has only the United States as an immediate neighbour, whereas the United States shares a border with Mexico from which, starting in the 1960s, it has obtained a remarkably large immigrant flow. The educational outcomes of the children of immigrants in the United States and Canada are distinct, with those in Canada attaining on average more years of schooling and higher skills, as measured by test scores, by the time they are adults.
To establish a relevant context, the chapter begins with a discussion of American and Canadian immigration history, which is reflected in the outcomes of immigrants’ offspring. It then turns to the substantial issues, addressing in turn the intergenerational transmission of educational, and labour market, outcomes.
American and Canadian immigration systems
“Traditional” immigration, before the 1960s
Immigration is foundational in both the American and Canadian national stories (Smith and Edmonston, 1997 for the United States; Green and Green, 2004 for Canada). Each experienced the so-called Great Migration of the late 1800s and early 1900s, but both countries’ immigration rates (i.e. immigration as a percentage of the population) were much reduced during the First World War. Subsequently, immigration continued until the Great Depression (about 1930) but was then essentially curtailed until the end of the Second World War.
After the Second World War (circa 1947), the broad strokes of the two nations’ immigration history diverge. Figure 7.1 illustrates how the immigration rates developed in Canada and the United States between 1940 and 2015. While Canada recorded high immigration rates after the war that were similar to those it experienced in the 1920s (although similar absolute numbers correspond to a lower rate given the intervening population growth), US immigration never returned to pre-Depression levels or rates. With one exception (discussed below), the US immigration rate increased very slowly from the late 1940s to the present. Borjas and Katz (2007) point out that in the 20th century the share of the US population that comprised immigrants peaked in 1910 at around 23% and then steadily declined to around 5% in 1970; at that point it started increasing again, reaching around 13% in the year 2000.
Figure 7.1. Immigration rate as a percentage of the population, United States and Canada, 1940-2015
Source: Canada: Immigration, Refugees and Citizenship Canada (IRCC, multiple years), Facts and Figures: United States: Department of Homeland Security (DHS, multiple years), Yearbook of Immigration Statistics.
Whereas post-war US immigration was very stable from year to year, Canada adopted a pro-cyclical immigration policy (Figure 7.2): increases during economic booms and marked decreases during recessions (with exceptions such as the mid-1950s Hungarian refugee flow). The country dropped this policy in the early 1990s.
Figure 7.2. Canadian unemployment and immigration rates (1950-2010)
Source: Canada – Immigration, Refugees and Citizenship Canada (multiple years), Facts and Figures; United States – Department of Homeland Security (multiple years), Yearbook of Immigration Statistics.
Importantly, the United States has had significant undocumented immigration flows since the 1960s, much of it from Mexico and to a lesser extent Central America. Baker and Rytina (2013) of the US Department of Homeland Security (DHS) estimate that as of 2012 there were about 11.4 million undocumented immigrants, up from 8.5 million in 2000. About 59% of this stock was born in Mexico, with at least an additional 14% from Central America. Although intrinsically hard to measure, Borjas (2017) suggests that as many as 30% of the total stock of US immigrants aged 20-65 in 2012/13 may be undocumented. There are no credible estimates of the undocumented population for Canada. However, a report from the Office of the Auditor General of Canada employing information from the country’s Ministry of Immigration, Refugees and Citizenship (IRCC) identified about 63 000 individuals in the late 2000s, although this was probably an underestimate of the total. Therefore, in the United States the undocumented represent around 30% of all immigrants (legal and undocumented), whereas in Canada they represent (at the upper bound) about 5%. The major US regularisation of the late 1980s and early 1990s – following from the 1986 Immigration Reform and Control Act – provided amnesty and a path to citizenship (Borjas, 2007). Even during that massive spike in legal immigration, as seen in Figure 7.1 when many undocumented immigrants became regularised, the US rate stayed below the legal immigration rate in Canada.
Of course, the US population is roughly 10 times that of Canada so it is important to keep absolute numbers in mind. The number of legal immigrants arriving in Canada each year is one-third to one-quarter of the US count.
“New” immigration, since the 1960s
Both the United States and Canada dramatically changed their immigrant selection processes in the mid- to late-1960s. Previously, the source countries for both had been virtually restricted to Europe and their North American neighbours. However, this was recognised as discriminatory, and the United States shifted to a system favouring family reunification/sponsorship, while Canada adopted a system with greater emphasis on economic migration and introduced the first immigration points system in 1967, increasing the range of source countries markedly. Both countries also managed major programmes offering citizenship to refugees.
Table 7.1 documents annual legal immigration by immigration class, averaging the years 1990 to 2015. Just over 60% of US immigrants were family members. In contrast, for Canada the family class averaged around 30%. Economic immigrants formed only about 13% of US immigration, with most sponsored by employers, whereas in Canada their share is just over 55% with, until recently, employers playing a much lesser role. This composition has implications with regard to immigrants’ children, since economic class immigration comprises a higher share of intact nuclear families than does the family class. This implies that Canada receives more child migrants who spend their formative years in the recipient country, and also has more individuals with two immigrant parents. Refugees are an interesting comparison that illustrates the confusion that can arise in popular discussion when rankings by absolute numbers and percentages are intermixed without sufficient explanation. As seen in Table 7.1, refugees represent roughly 12% of the legal permanent migration flow in both countries, but on average the United States accepts just over 120 000 annually, compared to Canada at just under 30 000. Conversely, the same refugee arrivals are approximately 0.04% of the US population, but about 0.09% of that of Canada. Different stories can be told by selectively quoting these statistics.
Table 7.1. Annual immigration by immigration class to the United States and Canada, averaging years 1990 to 2015
Note: Economic in Canada includes accompanying family members of economic migrants.
Source: Canada – RCC, multiple years; United States – DHS, multiple years.
Table 7.2. Top ten source countries by resident stock, 1960/61 and 2011
Sources: Panel 1: 1960 US Census of Population, and 1961 Canadian Census of Population. Panel 2: 2011 Canadian National Household Survey and 2007-2011 American Community Survey, U.S. Census Bureau.
Immigration subsequent to the 1960s reforms is sometimes called the “new” immigration. For both countries immigrant source countries globalised, with greater numbers coming from Africa and especially Asia. Also and notably, the United States (but not Canada) saw a marked increase in immigration from Central America, and a dramatic increase from Mexico. Table 7.2 illustrates this shift. Panel 1 summarises the stock (i.e. the number of immigrants resident) and the distribution of immigrants by source country in 1960/61; panel 2 does the same around 2011. There is clearly an important shift in source countries. Increasing diversity, except for Mexican immigration in the United States, is also indicated by the declining share of immigrants represented by the top 10 countries as seen by comparing the ‘total’ rows.
Immigration history’s implications for analyses across the immigrant generations
The children of immigrants are influenced by their parents’ characteristics and outcomes, especially economic and labour market outcomes, by the cultural characteristics of the community in which they reside, and by characteristics of the surrounding society, especially those of the education system and labour market. Of particular relevance for both countries is a well-documented (and appreciable) decline in earnings at entry across successive cohorts of new immigrants that began following the onset of the new migration. (See for example Borjas, 1985, 1995 and 2015 for the United States, and Baker and Benjamin, 1994; Aydemir and Skuterud, 2005; and Picot and Sweetman, 2005 and 2012 for Canada.)
Immigration from Mexico – as well as Hispanic immigration from Central and South America – is also an important issue in the United States (Haller, Portes and Lynch, 2011a, b; Alba, Kasinitz and Waters, 2011). Borjas (2007, p. 1) points to the “heated [US] debate over the possibility that Mexican immigrants and their descendants may assimilate slowly – relative to the experience of other immigrant waves and this slow assimilation may lead to the creation of a new underclass.” Compounding this is the large share of Mexican immigrants who are undocumented, and therefore frequently do not have full access to social and health services even though children of undocumented immigrants born in the United States are citizens themselves. For Canada, no one source country similarly dominates.
Table 7.3 presents the geographic origins of the stocks of both immigrants and the children of immigrants around the year 2000. Despite the “children” label, all of those represented in this table are adults. Also, although some of the younger individuals in the “children of immigrants” category may be young enough to be the children of the older individuals in the “immigrant” category, there is no parent-child relationship between these columns. While neither column exclusively comprises traditional or new source countries (roughly pre- or post-1970), for both the United States and Canada the children of immigrants have parents who were drawn more heavily from the so-called traditional source countries of, primarily, north-western, eastern and southern Europe, and the English-speaking developed world whereas the immigrant column reflects the new immigration to a greater degree. For example, while Asia comprises 20% to 30% of newer immigrants, only between 2% and 5% of the children of immigrants originate from this region.
Table 7.3. Geographic origins of immigrants and the parents of native-born children of immigrants around 2000
Source: Aydemir and Sweetman (2008), Table 3. Data: US 1998-2004 CPS; Canada 2001 Census.
Table 7.4 presents shares of adult and child immigrants, the children of immigrants, and the children of the native born, in Canada and the United States around 2000 and 2012. Perhaps surprisingly, the immigrant share of the population in Canada is only slightly over 50% higher than that in the United States, whereas Figure 7.1 and Table 7.1 suggest a larger difference. Two major factors explaining (part of) this are, first, the larger number of undocumented immigrants living in the United States who are captured in surveys of the domestic population; and second, emigration from Canada, which is considerable (see e.g. Aydemir and Robinson, 2008). While most immigrants arrive in the United States when they are older than 11 or 12, a higher share arrives at young ages in Canada. In contrast to the modest difference in immigrant population shares, the population share of the children of immigrants is almost 2.4 times larger in Canada. This is more in line with expectations and also reflects the earlier historical period, the decades following World War II, when the gap in immigration as a proportion of each country’s population was larger. As expected given the above-mentioned differences in family/economic class flows, the percentage of native-born children with two immigrant parents is markedly higher in Canada.
In the datasets employed for most contemporary analyses, virtually all of the descendants of the Great Migration (i.e. migration in the late 1800s and early 1900s) and earlier are primarily of European ancestry and are in the comparison group labelled children of the native born. Further, studies of the children of immigrants using older data focus primarily on individuals whose parents immigrated prior to the immigration reforms of the 1960s, whereas such studies using more recent data capture a higher share of members of the “new” immigration whose parents immigrated after the 1960s reforms. It is clear that findings for immigrants or their children based on immigration prior to the 1960s are not necessarily transferable to subsequent cohorts, as evidenced by the decline in labour market outcomes of new immigrants starting in the 1970s. Differences between the children of the old and new migration may exist as a result of changing ethnicity, parents’ labour market success, and other similar factors.
Since in much empirical research the grandchildren from of the Great Migration and earlier arrivals form the comparison group in analyses of earnings and other outcome variables, it is important to understand that the compositions of these population subgroups in the United States and Canada are very different, as seen in Table 7.3. The share of the children of the native born that is a visible minority is much higher in the United States. In large part this reflects the large African-American US minority (the Hispanic community is normally classified as non-visible minority). In Canada, the largest visible minority group among those with the native born parents comprises Indigenous peoples. Both these populations have faced discriminatory educational and labour market constraints. Some research uses the entire group of children of the native born as the comparison group, whereas other work focuses on the non-visible minority subset. That these two approaches yield different empirical results is not surprising; those reading the literature, therefore, need consider how the data for analysis are selected when interpreting results.
One crucial measurement issue in studying intergenerational issues related to immigration is the complexity associated with measuring later generational membership. Duncan and Trejo (2014, 2017) point out that ethnic identification becomes increasingly problematic for the children, grandchildren and subsequent generations of immigrants given the complexities of intermarriage and self-identification. Moreover, the authors estimate that the rate of ethnic “stickiness” varies systematically across ethnic groups, with Hispanic immigrants more likely to intermarry and cease reporting Hispanic ethnicity than other immigrant ethnic groups. In particular, high-income/high-education individuals from some of the largest Hispanic source countries are more likely to intermarry, and – by the time their family’s immigration experience dates back to their grandparents – are less likely to self-identify as Hispanic than immigrants from other groups. This biases estimates of intergenerational integration and makes Hispanic, and especially Mexican, immigrants appear to integrate more slowly. This is not to say that results are overturned, but the gaps observed in studies using self-reports of ethnicity as a basis for identification may be too pessimistic regarding Hispanic integration.
Table 7.4. Generational population shares of immigrants in the United States and Canada, 2000 and 2012, percentages
Note: Adult immigrants are defined as arriving aged 12 or older in 2000, and 11 and older in 2012, and child immigrants are those arriving younger than those ages.
Sources: 2000: Aydemir and Sweetman (2008), Tables 1 and 2. US Bureau of Labor Statistics, Current Population Survey microdata, 1998-2004; Statistics Canada, Canada 2001 Census microdata. Ages 25-65.
2012: PIAAC 2012; public use microdata. Ages 25-65.
Educational and labour market outcomes of the children of immigrants
Overviews related to the children of immigrants in the United States and Canada are by Aydemir and Sweetman (2008), Picot and Hou (2011a, b) and Sweetman and van Ours (2015). Card, DiNardo, and Estes (2000) provide a thorough analysis of the US context from the 1940s to the 1990s while Chiswick and DebBurman (2004), using data from 1995, find that the children of immigrants in the United States have higher levels of education than both the foreign-born and the children of the native born. Feliciano and Lanuza (2017) survey and interpret the US literature on immigrant intergenerational educational attainment.
A demographic overview of the children of immigrants in the United States is provided by Taylor et al. (2013) using data from 2002. They show that the children of immigrants, of whom 36% have graduated from college, are 5% more likely to have done so than the average American (without controlling for age; the children of immigrants are about eight years younger on average). Household income is almost exactly the national average as is homeownership, but the poverty level of immigrants’ offspring is slightly below the national average. Ethnically, in 2002 the children of immigrants looked very different from the children of the native born: the composition of the children of immigrants was 46% white, 4% black, 35% Hispanic, 12% Asian, and a small “Other” grouping. For the children of non-immigrants in contrast, the composition was 78% white, 13% black, 6% Hispanic, and < 0.5% Asian. For Canada, basic descriptive statistics are provided by Aydemir and Sweetman (2008) and Picot and Hou (2011a) using data from 2000. The children of immigrants, just over 23% of whom have graduated from university, are about 6.5% more likely to have done so than the average child with native born parents (without controlling for age; this group is about the same age). Annual earnings are very similar to or slightly higher than the national average.
Educational attainment
There are noticeable differences between the United States and Canada in the educational outcomes of the children of immigrants that mirror the differences of the immigrants’ outcomes. Table 7.5 explores these using the OECD Programme for the International Assessment of Adult Competencies (PIAAC) data, which are recent and employ a consistent sampling frame across countries. The table presents regression results estimating years of schooling and controlling only for age. The sample size for Canada is appreciably larger, providing more precise estimates. The intercept (i.e. constant) reflects the average years of schooling of the children of the native born, including visible minorities, at age 40. Each generational coefficient represents a difference relative to that group.
Table 7.5. Estimated years of schooling for adult and child immigrants, the children of immigrants and the children of the native born, United States and Canada, 2012
Notes: Individuals aged 25-69 with valid information on education, age at immigration, and parental and own place of birth. Controls are (age – 40) and (age – 40)2. Immigrants that arrived before the age of 12 are regarded as immigrating children. Heteroskedasticity consistent standard errors are in parentheses: * p<0.05, ** p<0.01, *** p<0.001. The constant is interpreted as the average years of schooling of a 40-year-old child of the native born, and other coefficients are differences from that average holding age constant. For example, a 40-year-old male immigrant in the United States is estimated to have 0.8 years less schooling than a 40-year-old man with no immigrant background.
Source: PIAAC, 2012. Authors’ calculations.
Focusing first on the children of the native born, females have slightly more schooling than males on average, but for each gender the United States and Canada are effectively identical. Large differences in the average level of education for immigrants are evident however, with those in the United States having substantially fewer years of schooling than the children of the native born in accord with US immigrant selection processes. In contrast, in accord with Canada’s immigration selection processes, immigrants have higher years of schooling than the children of the native born. This has appreciable ramifications for both young immigrants and the children of immigrants. Holding intergenerational correlations constant, Aydemir and Sweetman (2008, Table 10) project future national educational levels given the immigrants’ education levels and their share of the population. They predict marked increases for Canada, and few changes from the current norms for the United States.
Although the standard errors are large, making inference difficult for the United States, there is no evidence that immigrating children are faring particularly well there. In contrast, in Canada the coefficient estimates for both genders are positive, statistically significant, and economically large. This is a dramatic difference that mirrors the difference for adult immigrants, but unlike the results for adult immigrants, the difference with immigrating children results from events and choices made within the host country. It reflects both the receptivity of the domestic education systems to these immigrants, and the aspirations of the immigrant parents and children. (See Raleigh and Kao, 2010 for more on parental aspirations across immigrant ethnic groups in the United States.)
On a positive note, results for the native born children of immigrants show them to be obtaining at least as many years of schooling, and sometimes appreciably more, than the children of the native born comparison group. In the United States, males obtain more years of schooling while females are indistinguishable from them, and for Canada both estimates are positive and statistically significant. The finding that male children of immigrants have a higher level of education than the children of the native born in the United States differs from that observed in some studies. For example, Picot and Hou (2011a) report the US children of immigrants as having average years of schooling comparable to that of the children of the native born. However, many of the studies they surveyed used the sub-group of non-visible minorities among the children of the native born as the reference group, who (as discussed) on average have better outcomes, whereas the sample in Table 7.5 includes the entire population of the children of the native born since ethnicity is not available in the data employed. This discrepancy may explain the difference. Also, the data underlying Table 7.5 are more recent and have a greater percentage of children of the “new immigration”.
Table 7.6 expands the ideas of Table 7.5 but splits the sample into three parts, based on each respondent’s parents’ level of education. These are defined using the parent with the highest level of education as: low, i.e. below high school or equivalent; medium, i.e. some or complete upper secondary and post-secondary non-tertiary education; and high, i.e. some or complete tertiary education. In all cases, on average more highly educated parents have more highly educated children. Using the children of the native born as a base case, those of highly educated parents have comparable years of schooling in Canada and the United States, but the children of native born parents with low levels of schooling have appreciably more years of education in Canada.
Focusing on the children of immigrants in Table 7.6, there are marked differences across both genders and countries. For males in the United States, children of highly educated immigrant parents outperform the children of the native born with parents in the same education category in terms of years of schooling. In contrast, there is effectively no difference for this group in Canada. At the other end of the spectrum, male children of immigrants whose parents have low levels of education attain more years of schooling relative to the children of the native born with the same level of education in Canada than in the United States. Women have very different patterns. In the United States there is effectively no difference between the children of immigrants and those of the native born, whereas the Canadian women somewhat reflect the case of the US men. Across the entire table, the intergenerational transmission of education seems weaker in Canada than the United States, and it is especially weak for immigrant children and the native born children of immigrants in Canada. For the children of immigrants in Canada, parents’ education matters far less in determining the final education level of their children than it does for the children of the native born, in that the children of immigrant parents with low levels of education do extremely well in the education system. As will be seen, this seems to result from actions taken on both the supply side (i.e. the school) and the demand side (i.e. immigrant parents and their children) of the equilibrium.
Beyond years of schooling, basic skills as measured by tests scores are a complementary educational outcome to years of schooling. To that end, regressions with literacy scores from the PIAAC as the dependent are presented in Table 7.7. Though not presented, regressions using numeracy scores have a pattern of results that is broadly similar. The regression specifications in Table 7.7 are similar to those in Table 7.5, except that the number of regressions is twice as large since versions controlling and not controlling for educational attainment are both presented. In all of the regressions the test scores are normalised to have mean zero and variance one, where the mean is the average of the two countries with each given equal weight. The coefficients on the intercepts, therefore, are interpreted as measuring standard deviations from the two-country mean for the children of the native born whose highest level of education is high school, and the coefficients on the regressors measure differences (in standard deviations) from that group.
Table 7.6. Intergenerational transmission of years of schooling by gender, United States and Canada, 2012
Notes: See Table 7.5. Each constant is interpreted as the average years of schooling of a 40-year-old child of a native born parent in the relevant educational category, and the other coefficients in the same column are differences from that average for those whose parents are in the same educational category. For example, a 40-year-old male immigrant in the United States whose parents attained only low education is expected to have 0.9 years less of schooling than a 40-year-old man without an immigration background whose parents also obtained only low education.
Source: PIAAC, 2012. Authors’ calculations.
Table 7.7. Proficiency in the PIAAC measure of literacy by immigrant generation by gender, United States and Canada, 2012
Notes: See Table 7.5. The omitted educational group are high school graduates.
Source: PIAAC, 2012. Authors’ calculations.
Oddly, while Table 7.5 shows male children of immigrants in the United States to have about three-quarters of a year of additional education compared to the children of the native born, no such increment in literacy test scores is observable in the first column of Table 7.7. Continuing to focus on the first regression of each set, which does not control for years of schooling, for female children of immigrants in the United States, there is no gap in years of education, and there is a marginal statistically significant negative gap in test scores. In contrast, both male and female children of immigrants in Canada are ahead, in terms of both years of education and test scores. For numeracy (not shown) the pattern is the same for males, but neither coefficient for females is statistically significant. Overall, it is unclear why the years of schooling for the children of immigrants in the United States are not translating into skills, as measured by these assessments, at the same rate as for the children of the native born.
When controls for education are introduced into the regressions in the second column for each country/gender, the coefficients for the children of immigrants in the US sample are not much affected, whereas those for the Canadian sample are reduced and become statistically insignificantly different from zero. That is to say, the children of immigrants in the United States appear to obtain a somewhat lower test score increment compared to education than do the children of the native born, whereas in Canada the additional education accrued by the children of immigrants appears to “explain” their higher average test scores. Despite this observation with respect to the children of immigrants, test scores overall are clearly correlated with years of schooling, as evidenced by the coefficients on the education variables.
Educational outcomes and observed characteristics
Observed characteristics are associated with differences in educational outcomes between immigrating children, the children of immigrants, and the children of the native born. Before turning to a detailed discussion of various findings, a brief overview is provided regarding three clusters of variables: parents’ education and income, ethnicity, and urbanisation. Parents’ education and income are found to be less important in determining educational outcomes for the children of immigrants than is the case among the children of the native born; this is particularly the case in Canada. Moreover, parental income is much less important than education. Ethnicity (sometimes proxied by source country), likely involving cultural factors operating at the community level, appears to play a very important role in determining educational attainment. Average levels of such attainment vary dramatically across ethnic groups, and this is particularly evident for the new immigration. Additionally, compared to the children of the native born, the children of immigrants are not only much more urbanised but also more likely to live in the very largest urban centres. Education levels also tend to be higher in more highly urbanised regions. Educational gaps between the children of immigrants and the children of the native born can, therefore, look very different if the comparison group comprises the entire population of the children of the native born, as opposed to only those residing in the same local labour market.
An older line of research addressing the intergenerational integration of immigrants predating the new migration focuses not on individuals but on aggregate ethnic/source country outcomes. Both Borjas (1994, for the United States) and Dicks and Sweetman (1999, for Canada) find important ethnic group effects that persist from one generation to the next with convergence taking on the order of four generations.
Research using microdata also finds strong ethnic/source country effects on outcomes for the children of immigrants. In particular, in the United States there is substantial literature addressing Hispanic, especially Mexican, immigration. Landale, McHale, and Booth (2010) and Fortuny and Chaudry (2011) document that in 2007-08, 43% of the children of immigrants in the United States were Hispanic, and that the percentage is increasing. A very old story is of Hispanic immigrants having low education, and of their children having similarly low education. However, research at least as long ago as Smith (2003), focusing on the broad Hispanic/Latino community, shows that though starting from a very low base, educational integration is faster than had been appreciated. Slightly less optimistically, Trejo (2003), focusing on the somewhat narrower category of Mexican immigrants, finds marked improvements in educational attainment between the immigrants and their children, but the process slows down appreciably after that. What Trejo observes between the immigrants and their children is a very substantial increase to an average that is near high school graduation and not too far from what might be expected given compulsory schooling ages. This highlights that a key challenge facing the children and subsequent generations of Mexican immigrants is postsecondary education (Park and Myers, 2010; Grogger and Trejo, 2002).
Fry (2002) observes that it is not so much access as completion that is at issue. Hispanic high school graduates in the 1990s enrolled in postsecondary education at rates comparable to those of other ethnic groups, but were more likely to enrol part-time, and in community colleges as opposed to four-year institutions. Many failed to complete their degrees. Also, the enrolment rate in graduate and professional education is markedly lower than that for other ethnic groups in the United States. Policies focusing on postsecondary persistence and completion are important results from Fry’s work. However, he notes, as do Duncan and Trejo (2014), that there are important differences within the Hispanic community. Especially, immigrants of Cuban origin have higher, and youth of Mexican and Puerto Rican origin have substantially lower, postsecondary completion rates than the Hispanic average.
Lopez (2009) directly surveys relevant Hispanic youth and finds that while 90% recognise postsecondary education as an important pathway to success, only about 48% plan on attending. The main reported impediment is financial pressure to support their family. Additional reasons included poor English skills, a dislike of school, and aspirations for employment in occupations that do not require postsecondary education. A disaggregation of these youths into immigrant children and the children of immigrants shows that the latter group is more likely to both plan to and actually attend postsecondary education. Also importantly, Perreira and Spees (2015) point out the substantial difficulties in attending postsecondary education among the large proportion of the postsecondary-aged Hispanic community who are undocumented. Plausibly, this has repercussions on ethnic-group-level social capital and aspirations that spill over to all youth in the community.
In terms of parental aspirations, Dondero and Humphries (2016) compare the college savings rates of Asian, Hispanic and white parents for their children in grade 10. They observed that Asian immigrants are more likely to save for their children’s education, and have greater savings, than native born parents (irrespective of ethnicity). In contrast, Hispanic immigrants are less likely to save than native born parents, but looking only at those who do save, they have roughly the same amount put aside.
Fry (2007) points out the challenges facing those learning English as a second language in the school system, and notes that fully 70% of this group is Hispanic, with many comprising the children of immigrants. He points out the relatively low scores of Hispanics and notes their particularly large language deficiencies. Akresh and Akresh (2011) pursue this issue further by using randomisation in the assignment of the language for achievement tests (either English or Spanish). They note that foreign-born children who arrive at an older age score higher when the test is offered in Spanish; those who arrive younger are, on average, indifferent to the language of the test; and the children of immigrants fare better when the test is offered in English. This suggests that integration towards English is occurring.
One reason sometimes put forward for the relatively slow integration of the Hispanic community into higher levels of education in the United States is the size of the community. In a sufficiently large community, English language acquisition is less important. As pointed out by Lazear (2007), shortly after arrival 80% of non-Mexican immigrants are fluent in English, in contrast to only 49% of Mexican immigrants. Moreover, time in the United States does not appreciably change this percentage and the share of Mexican immigrants who are fluent in English, which hits its maximum around 20 years after migration, is 60%. Lazear also documents larger-than-average enclaves among Mexican immigrants, and some argue that return migration because of the proximity of Mexico also plays an important role.
Borjas (2015) pushes Lazear’s argument further. Although focusing on immigrants and not their children, he finds evidence suggesting that the rate of English language acquisition in the United States is significantly slower for larger non-English speaking groups. In particular, the substantial size of Hispanic enclaves reduces the incentive for resident populations to learn English, which reduces their long-term economic success. Related to this, Potochnick (2014) studies grade 10 math and reading test scores for the children of immigrants across three categories of US states: established immigrant recipient states, new immigrant recipient states, and those in between. She notes that immigrants have superior outcomes in the new immigrant recipient states where there are no well-established enclaves.
A substantial literature addresses approaches to improving educational outcomes among students from disadvantaged backgrounds, especially minority ethnic group students, in the United States (see e.g. Betts, 2011). The findings within this literature are mixed, and some studies have weak research designs. One recent contribution with a stronger design is by Card and Giuliano (2016), who study within-school streaming of high-ability students in grade 5. They find marked improvements in test score outcomes for gifted and high-achieving minority students who are clustered together in high-ability classrooms, with no evidence of negative spillovers on nonparticipants. Overall, the benefits appear to be concentrated among black and Hispanic high-ability and gifted children.
While streaming high ability students may be beneficial, there is also evidence of problems from streaming into low ability classes too early for the children of immigrants where the parents are foreign language speakers, especially if it is difficult to subsequently change streams. Worswick (2009) points out that in the Canadian context, the children of immigrants have low performance in vocabulary at age 6, but then accelerate and by age 14 have outcomes comparable to the children of Canadian-born parents. There is some evidence that if these children are streamed too young, they will not have had sufficient opportunities to catch up, meaning that the streaming decision is partly driven by the language spoken at home. (See Sweetman and van Ours (2015) for a discussion of these issues.) More broadly, many perceive that one of the advantages of the Canadian education system for immigrating children and the children of immigrants is that it provides numerous opportunities to “bounce back” (Finnie, Laporte and Sweetman, 2010). This may account for some of the positive outcomes for the children of immigrants seen in Table 7.5 through Table 7.7.
In Canada substantial heterogeneity in educational attainment is observed across ethnic groups, but the ethnic distribution is very different from that in the United States and immigrant parents have much higher levels of postsecondary attendance. Using longitudinal survey data whose starting point was an early Programme for International Student Assessment (PISA), Childs, Finnie, and Mueller (2017) observe that university participation rates by age 21 for immigrating children, the children of immigrants and the children of the native born are, respectively, 57%, 54% and 38%. This sample of the children of immigrants are those of the “new” immigration to Canada discussed by Finnie and Mueller (2010).
Figure 7.3 through Figure 7.6 reproduce results from Childs, Finnie, and Mueller (2017) for Canada. The figures represent predictions from a model holding observable characteristics constant. Each displays the probability of university access by source region for the children of immigrants, and for comparison the children of the native born (labelled as the children of non-immigrants). The key difference in each of the plots is the x-axis, which in turn depicts family income; years of schooling for the parent with the most education; average high school grades; and PISA reading score. As will be seen, many of these traditional correlates of postsecondary access do not carry over to many immigrant source country groups – though some, especially the so-called “Anglosphere” (e.g. United States, United Kingdom, Australia, etc.) are similar to the non-immigrant group. Note that the United States is included in the “Anglosphere” as opposed to “the Americas”, with the latter including Mexico, Central and South America and the Caribbean.
Figure 7.3 suggests that, on average, postsecondary attendance is relatively insensitive to family income for the children of immigrants in many ethnic groups, and where there is a very strong relationship, for parents from Africa and China, the mean level of attendance is quite high for the lowest income category. Most remarkable is that among every ethnic group, with the exception of those whose parents were from “the Americas”, children of immigrants have a higher probability of attending university than the children of the native born (i.e., children of non-immigrants).
Figure 7.3. Children of immigrants' probability of university access by parental source region and family income
Note: Predictions from a regression model.
Source: Data from Childs, Finnie and Mueller, 2017.
A similar exercise is presented in Figure 7.4, which shows parents’ years of schooling. A general upward trend is observed for most but not all ethnic groups, but it is not nearly as strong for those ethnic groups with high access rates as it is for the children of the native born.
Figure 7.4. Children of immigrants' probability of university access by parental source region and years of schooling
Note: Predictions from a regression model.
Source: Data from Childs, Finnie and Mueller, 2017.
Figure 7.5 turns to an intermediate outcome: the final average high school grade. The key insight here is that ethnic groups most linked with postsecondary attendance seem to gain access even when they have relatively low marks. There is a gap in university attendance between the children of the native born and the children of immigrants for most (but not all) ethnic groups among those with high grade point averages, but the gap is substantially larger in the middle or bottom of the distribution.
Figure 7.5. Children of immigrants' probability of university access by parental source region and students' high school grades
Note: Predictions from a regression model.
Source: Data from Childs, Finnie and Mueller, 2017.
A conceptually similar plot is presented in Figure 7.6, although this looks at the PISA reading score at age 15. The children of immigrants in many ethnic groups have high probabilities of attending university despite relatively low test scores.
Although Figures 7.3-7.6 are descriptive, cultural or ethnic determinants of access appear to be indeed large; perhaps that has to do with the ethnic human capital discussed by Borjas (1992), Postepska (2017), and Finnie, Mueller and Sweetman (2016). The traditional factors depicted in these figures do not appear to be nearly as important for these “new” children of immigrants as they are for the children of the native born. Moreover, the relevant educational aspirations appear to be in place at a very young age. Aydemir, Chen, and Corak (2013) point out the much weaker relationship between parents’ and child’s education for immigrants in Canada compared to non-immigrants. Focusing on ethnic group averages, Luthra and Soehl (2015) make a similar observation for the children of the “new” immigration in the United States. Of course, the relative sizes of ethnic groups vary dramatically across the two countries. Baum and Flores (2011) agree but also point to important source country groups with low average levels of post-secondary access, and suggest that it is necessary to understand specific differences between ethnic groups. Feliciano and Lanuza (2017) survey ethnic group level differences in educational attainment among the children of immigrants in the United States.
Figure 7.6. Children of immigrants' probability of university access by parental source region and students' reading score
Note: Predictions from a regression model.
Source: Data from Childs, Finnie and Mueller, 2017.
On a different note, the relationship between the educational attainment of immigrating children and the native-born children of immigrants, as a function of their parents’ immigration classification, is studied by Hou and Bonikowski (2016). The authors observe substantial differences in educational outcomes of the children of immigrants across immigrant entry categories. Children whose parents are in the skilled worker (points system) and business categories had university completion rates that were about two and a half times higher than those of individuals whose parents entered in the family class. Refugees were somewhere in between, but closer to the family class. Consistent with the broad finding by, for example, Bleakley and Chin (2004, 2008, 2010, for the United States) and Schaafsma and Sweetman (2001, for Canada) who observe that parental influences matter less for children who arrived very young. Hou and Bonikowska observe that the parents’ admission class matters less for those who arrive in Canada at preschool age. Controlling for parents’ characteristics as observed at arrival reduces the gaps by less than 50% for the family class, and by as much as two-thirds for refugees. Focusing on immigrating children, Evans and Fitzgerald (2017) find that child immigrants who enter the United States before age 14 have similar educational outcomes as the children of the native born, whereas those who enter at a later age have poorer outcomes that are largely attributable to language skills.
Labour market outcomes
Understanding the data used to study the intergenerational economic integration of immigrants into host countries is crucial for interpretation. Much research in the United States and Canada looking at labour market outcomes among immigrants focuses on annual employment earnings, probably because many studies rely on census data where this outcome is reported. However, some studies use data sources that contain hourly wages. The difference is significant since immigrants tend to work longer hours than do the native-born. Also, many studies restrict the sample for analysis to the (full-time) employed, which misses differential employment levels. Poverty is sometimes chosen as an alternative outcome measure.
Broadly speaking, in both countries the labour force participation rates and the unemployment rates of immigrants who arrived as children, and both the children of immigrants and the native born, are comparable (Picot and Hou 2011b). However, as pointed out by Borjas (2017), adult immigrants (especially those undocumented) in the US are distinct in having higher labour force participation rates and longer working hours. Moreover, immigrants commence working younger and retire later.
In interpreting the earnings of the children of immigrants relative to the children of the native born for both the United States and Canada, it is important to distinguish between unadjusted results (or results conditional only on age) and those that are conditional on an array of observed characteristics. Unconditionally, in both countries the children of immigrants of both genders have earnings that are equal to or greater than those of the children of non-immigrants (Card, DiNardo and Estes, 2000; Picot and Hou, 2011b; Aydemir and Sweetman, 2008). In the United States the earnings of children of immigrants are close to the same as those of the children of the native born (if one compares individuals of the same age there is effectively no gap). In Canada the children of immigrants have an earnings advantage on the order of 10-15% (those with two immigrant parents tend to be on the higher side of the range), and are also more likely to be employed in professional occupations, than the children of the native born. Immigrating children in Canada have an earnings advantage, close to 20%, relative to the children of the native born; however, in the United States they have a 20% disadvantage for males and no gap for females (Aydemir and Sweetman, 2008) – broadly consistent with the education outcomes seen in Tables 7.5-7.7.
When regression controls (especially education, ethnicity and urban/rural residency) are taken into consideration (Aydemir and Sweetman, 2008; Picot and Hou, 2011b), the coefficients for the children of immigrants change only modestly for the United States. Unconditionally, the Mexican community has lower average annual or weekly earnings than the children of the native born, but conditional on observed characteristics – in particular education – it has higher, not lower, earnings. In part this is attributable to higher hours of work. In Canada, the story is very different. The substantial earnings premiums of immigrating children and the children of immigrants, found without controlling for observed characteristics, are not only reduced but the gaps also shift from positive to negative. Unconditionally, immigrating children and the children of immigrants have a substantial advantage, but – conditional on (especially) their higher level of education and urban residency – those premiums become deficits. Immigrating children and the children of immigrants are, on average, more highly educated and much more likely to live in a small number of large urban locations, both of which characteristics are associated with appreciable earnings premiums. Picot and Hou (2011b) point out that much of the negative gap is concentrated among the visible minority, and particularly the black, community of immigrating children and the children of immigrants. Also, as seen in Figure 7.5 and Figure 7.6, the children of immigrants are more likely to access higher levels of education with lower levels of high school marks and test scores.
Taking a longer-term perspective using US data from 1940 to 1996, Card, DiNardo, and Estes (2000) observe that despite the substantial shift in source countries associated with the policy reforms of the 1960s, the rate of intergenerational integration in earnings has changed little. They find that the children of immigrants, on average, continue to have earnings that close 50‑60% of the gap experienced by their fathers’ ethnic group. In Canada, there is only a weak relationship between the earnings of immigrants and those of their children (Aydemir, Chen, and Corak, 2009), with a higher percentage of the gap closing.
One interesting Canadian study by Skuterud (2010) attempts to sort out immigration status from ethnicity, since there are debates about whether the immigrant earnings gaps at arrival that have opened since the 1970s are primarily the result of ethnic discrimination. Consistent with earlier work by Schaafsma and Sweetman (2001) looking at only immigrants arriving during childhood, Skuterud observes that most of the immigrant intergenerational improvement in earnings occurs for the visible minority community and not for the white one. The largest decrease in the size of the gap occurs between adult immigrants and immigrant children or the children of immigrants. That is, intergenerational earnings growth is primarily a visible minority phenomenon in Canada. However, the gap does not entirely close. Even for the children of native born visible minority workers, there remains an earnings gap compared to the children of native born non-visible minority workers, albeit small compared to that observed for immigrants.
A much smaller branch of literature but one with tremendous potential for future research focuses on employer and/or workplace effects. It is motivated by research such as that by Corak and Piraino (2011), which finds substantial commonality in employers of fathers and sons (independent of immigration status). Although focusing on immigrants, Aydemir and Skuterud (2008) use matched employer-employee data and observe that male immigrants are non-randomly sorted across employers within cities and geographic regions. Further, earnings differences across employers substantially dominate those within employers in explaining the immigrant earnings deficit. In contrast, for female immigrants, within-establishment wage differentials appear to play a larger role. It is unclear to what extent this employer effect matters for the children of immigrants, but research suggests that the effects on their children are plausible and should be studied in future.
Conclusion
As it concerns an increasingly large segment of society, understanding the intergenerational integration of immigrants into North America’s society and economy is becoming more important. The children of immigrants’ share of the population and the workforce is projected to increase in both the United States (Taylor et al., 2013) and Canada (Morency, Malenfant, and MacIsaac, 2017). Continually shifting source countries is also an important issue.
The United States and Canada have clearly taken different paths – in part because of ongoing policy decisions, many of which commenced in the 1960s; in part because of geographic location; and in part because of global reputations and political/familial links. The United States has a much higher share of immigrants with low levels of education than does Canada, and this has important ramifications for both immigrating children and the children of immigrants. Interestingly, the children of immigrants in the United States who have parents with low levels of education seem to have outcomes comparable to American children of the native born with similarly educated parents. In contrast, in Canada the children of immigrants who have parents with low levels of education acquire education in excess of that for children of the native born who also have parents with low levels of education. The story of immigrants’ children in Canada is not only about the children of highly educated parents obtaining substantial education, but also about the children of immigrants with low levels of education obtaining substantial education.
Looking at the children of immigrants by immigration class shows that the children of refugees have better outcomes than those of family class immigrants, and both have poorer outcomes than the children of economic immigrants. However, the younger a child arrives in Canada, the less important are the immigration classes for predicting educational outcomes.
In the United States, on average, the children of immigrants have earnings that are indistinguishable from those of the children of the native born, and conditioning on the standard set of variables employed in these types of analyses (age, education, location and ethnicity) does not substantially alter that conclusion. Of particular note in the United States is the large Hispanic, especially Mexican, community of immigrants and their children. These immigrants have low average levels of education, and earnings that are relatively good conditional on that education but low unconditionally. There is a substantial increase in education and earnings for the children of Mexican immigrants relative to that of their parents; educational attainment increases to close to the end of high school. Progress beyond this point is slower. Several rationales have been investigated including the magnitude of the Hispanic community, which is argued to make English comprehension less valuable in the short run and so prevent economic progress in the longer. Another key rationale is a classic liquidity trap whereby individuals cannot invest in education because they need to support their families. Experiments, such as providing access to advanced study classrooms to gifted and high-achieving minority students, are being attempted in certain locations to address this issue.
In Canada the story is quite different since immigrating children and the children of immigrants have an earnings premium unconditionally. This premium is “explained”, and even reversed, once observed characteristics, especially education and geographic location, are taken into account.
Much work remains to be done in both countries to gain a better understanding of the intergenerational integration of immigrants. More work such as that by Card and Giuliano (2016) looking at efforts to assist disadvantaged students, many of whom immigrated as children or are the children of immigrants, can help to plot a course forward. Also, increased research looking at the role of firms in the intergenerational integration of immigrants may provide insights to assist in accelerating the integration process.
References
Akresh, R. and I. Redstone Akresh (2011), “Using Achievement Tests to Measure Language Assimilation and Language Bias among the Children of Immigrants.” Journal of Human Resources, Vol. 46/3, pp. 647–67. doi:10.1353/jhr.2011.0018.
Alba, R.D., P. Kasinitz and M.C. Waters (2011), “The Kids Are (Mostly) Alright: Second-Generation Assimilation: Comments on Haller, Portes and Lynch.” Social Forces, Vol. 89/3, pp. 763–74. doi:10.1353/sof.2011.0024.
Aydemir, A., W.H. Chen and M. Corak (2009), “Intergenerational Earnings Mobility among the Children of Canadian Immigrants.” The Review of Economics and Statistics Vol. 91/2, pp. 377–97.
Aydemir, A., W.H. Chen and M. Corak (2013), “Intergenerational Education Mobility among the Children of Canadian Immigrants.” Canadian Public Policy XXXIX (Supplement/Numéro Spécial). doi:10.3138/CPP.39.Supplement1.S81.
Aydemir, A. and C. Robinson. (2008), “Global Labour Markets, Return, and Onward Migration.” Canadian Journal of Economics, Vol. 41/4, pp. 1285–1311.
Aydemir, A. and M. Skuterud (2005), “Explaining the Deteriorating Entry Earnings of Canada’s Immigrant Cohorts, 1966-2000.” Canadian Journal of Economics, Vol. 38/2, pp. 641–71.
Aydemir, A. and M. Skuterud (2008), “The Immigrant Wage Differential Within and Across Establishments.” Industrial and Labor Relations Review, Vol. 61/3, pp. 334–52. doi:http://digitalcommons.ilr.cornell.edu/ilrreview/.
Aydemir, A. and A. Sweetman (2008), “First- and Second-Generation Immigrant Educational Attainment and Labor Market Outcomes: A Comparison of the United States and Canada.” Research in Labor Economics, Vol. 27/2298, pp. 215–70. doi:10.1016/S0147-9121(07)00006-4.
Baker, B.C. and N. Rytina (2013), “Estimates of the Unauthorized Immigrant Population Residing in the United States: January 2010.” (US) Department of Homeland Security, Office of Immigration Statistics, January, pp. 1–7. doi:http://www.dhs.gov/xlibrary/assets/statistics/publications/ois_ill_pe_2010.pdf.
Baker, M., and D. Benjamin (1994), “The Performance of Immigrants in the Canadian Labor Market.” Journal of Labor Economics, Vol. 12/3, pp. 369–405.
Baum, S. and S.M. Flores (2011), “Higher Education and Children in Immigrant Families.” The Future of Children, Vol. 21/1, pp. 171–93. doi:10.2307/41229016.
Betts, J.R. (2011), The Economics of Tracking in Education. Handbook of the Economics of Education. Vol. 3. doi:10.1016/B978-0-444-53429-3.00007-7.
Bleakley, H. and A. Chin (2004), “Language Skills and Earnings: Evidence from Childhood Immigrants.” Review of Economics and Statistics, Vol. 86, pp. 481–96.
Bleakley, H. and A. Chin (2008), “What Holds Back the Second Generation? Transmission of Language Human Capital among Immigrants.” Journal of Human Resources, Vol. 43/2, pp. 267–98. http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2440707&tool=pmcentrez&rendertype=abstract.
Bleakley, H. and A. Chin (2010), “Age at Arrival, English Proficiency, and Social Assimilation among U.S. Immigrants.” American Economic Journal: Applied Economics, Vol. 2/1, pp. 165–92. doi:10.1257/app.2.1.165.
Borjas, G.J. (1985), “Assimilation, Changes in Cohort Quality, and the Earnings of Immigrants.” Journal of Labor Economics, Vol. 3, pp. 463–89. doi:10.1086/298065.
Borjas, G.J. (1995), “Assimilation and Changes in Cohort Quality Revisited: What Happened to Immigrant Earnings in the 1980s?” Journal of Labor Economics, Vol. 13, pp. 201–45.
Borjas, G.J. (1992), “Ethnic Capital and Intergenerational Mobility.” Quarterly Journal of Economics, Vol. 107/1, pp. 123–50.
Borjas, G.J. (1994), “Long-Run Convergence of Ethnic Skills Differentials: The Children and Grandchildren of the Great Migration.” Industrial and Labor Relations Review, Vol. 47/4, pp. 553–73.
Borjas, G.J. (2015), “The Slowdown in the Economic Assimilation of Immigrants: Aging and Cohort Effects Revisited Again.” Journal of Human Capital, previously NBER Working Paper w19116, Vol. 9/4, pp. 483–518.
Borjas, G.J. (2017), “The Labor Supply of Undocumented Immigrants.” Labour Economics, Elsevier, Vol. 46 (February), pp. 1–13. doi:10.1016/j.labeco.2017.02.004.
Borjas, G.J. and L.F Katz (2007), “The Evolution of the Mexican-Born Workforce in the United States.” In Mexican Immigration to the United States, pp. 13–55. doi:10.1017/CBO9781107415324.004.
Card, D., J. DiNardo and E. Estes (2000), “The More Things Change: Immigrants Chidren of Immigrants in the 1940s, 1970s, and the 1990s.” In Issues in the Economics of Immigration, G.J Borjas (ed.). University of Chicago Press.
Card, D. and L. Giuliano (2016), “Can Tracking Raise the Test Scores of High-Ability Minority Students?” American Economic Review, Vol. 106/10, pp. 2783–2816. doi:10.1257/aer.20150484.
Childs, S., R. Finnie and R.E. Mueller (2017), “Why Do So Many Children of Immigrants Attend University? Evidence for Canada.” Journal of International Migration and Integration, Vol. 18/1, pp. 1–28. doi:10.1007/s12134-015-0447-8.
Chiswick, B.R. and N. DebBurman (2004), “Educational Attainment: Analysis by Immigrant Generation.” Economics of Education Review, Vol. 23/4, pp. 361–79. doi:10.1016/j.econedurev.2003.09.002.
Corak, M. and P. Piraino (2011), “The Intergenerational Transmission of Employers.” Journal of Labor Economics, Vol. 29/1, pp. 37–68.
Department of Homeland Security (multiple years), Yearbook of Immigration Statistics, US Department of Homeland Security.
Dicks, G. and A. Sweetman (1999), “Education and Ethnicity in Canada: An Intergenerational Perspective.” Journal of Human Resources, Vol. 34, pp. 668–96.
Dondero, M. and M. Humphries (2016), “Planning for the American Dream: The College-Savings Behavior of Asian and Latino Foreign-Born Parents in the United States.” Population Research and Policy Review, Vol. 35/6, pp. 791–823. doi:10.1007/s11113-016-9409-x.
Duncan, B. and S.J. Trejo (2017), “The Complexity of Immigrant Generations: Integration of Hispanics and Asians.” Industrial and Labor Relations Review, IZA Discussion Paper No. 6276.
Duncan, B. and S.J. Trejo (2014), “Assessing the Socioeconomic Mobility and Integration of U.S. Immigrants and Their Descendants.” The ANNALS of the American Academy of Political and Social Science, Vol. 657/1, pp. 108–35. doi:10.1177/0002716214548396.
Evans, W.N. and D. Fitzgerald (2017), “The Economic and Social Outcomes of Refugees in the United States.” NBER Working Paper 23498.
Feliciano, C. and Y.R. Lanuza (2017), “An Immigrant Paradox? Contextual Attainment and Intergenerational Educational Mobility.” American Sociological Review, Vol. 82/1. doi:10.1177/0003122416684777.
Finnie, R., C. Laporte and A. Sweetman (2010), “Dropping Out and Bouncing Back: New Evidence on High School Dynamics in Canada.” In Pursuing Higher Education in Canada: Economic, Social and Policy Dimensions, R. Finnie, M. Frenette, R.E Mueller and A. Sweetman (eds.). McGill-Queen’s University Press, Montréal and Kingston.
Finnie, R. and R.E. Mueller (2010), “They Came, They Saw, They Enrolled: Access to Post-Secondary Education by the Children of Canadian Immigrants.” In Pursuing Higher Education in Canada: Economic, Social, and Policy Dimensions, R. Finnie, M. Frenette, R.E. Mueller and A. Sweetman (eds.). Queen’s Policy Studies Series. McGill-Queen’s University Press, Montreal and Kingston.
Finnie, R., R.E. Mueller and A. Sweetman (2016), “The Cultural Determinants of Access to Post-Secondary (Higher) Education in Canada: Empirical Evidence and Policy Implications.” In Access and Expansion Post-Massification, B. Jongbloed and H. Vossensteyn (eds.). Routledge, London.
Fortuny, K. and A. Chaudry (2011), “Children of Immigrants: Immigration Trends.” Urban Institutem Brief No. 5. /citations?view_op=view_citation&continue=/scholar%3Fhl%3Den%26start%3D20%26as_sdt%3D0,33%26scilib%3D2&citilm=1&citation_for_view=0_lM5hsAAAAJ:pqnbT2bcN3wC&hl=en&oi=p.
Fry, R. (2002), “Latinos in Higher Education: Many Enroll, Too Few Graduate.” Pew Hispanic Center Report, p. 32.
Fry, R. (2007), “How Far behind in Math and Reading Are English Language Learners?” Pew Hispanic Center, pp. 1–32.
Green, A.G and D. Green (2004), “The Goals of Canada’s Immigration Policy: A Historical Perspective.” Canadian Journal of Urban Research, Vol. 13/1, pp. 102–39.
Grogger, J. and S.J. Trejo (2002). Falling Behind or Moving Up? The Intergenerational Progress of Mexican Americans. Public Policy Institute of California. San Francisco, CA.
Haller, W., A. Portes and S. Lynch (2011a), “On the Dangers of Rosy Lenses: Reply to Alba, Kasinitz and Waters.” Social Forces, Vol. 89/3, pp. 775–82. doi:10.1353/sof.2011.0018.
Haller, W., A. Portes and S.M. Lynch (2011b), “Dreams Fulfilled, Dreams Shattered: Determinants of Segmented Assimilation in the Second Generation.” Social Forces, Vol. 89/3, pp. 733–62. doi:10.1353/sof.2011.0003.
Hou, F. and A. Bonikowski (2016), “Educational and Labour Market Outcomes of Childhood Immigrants by Admission Class.” Statistics Canada, Analytical Studies Branch Research Paper, Vol. 377, p. 43.
IRCC (n.d.), Facts and Figures, Immigration, Refugees and Citizenship Canada.
Landale, N.S., S.M. McHale and A. Booth (2010), “Growing Up Hispanic: Health and Development of Children of Immigrants.” Urban Institute, Fact Sheet.
Lazear, E.P. (2007), “Mexican Assimilation in the United States.” In Mexican Immigration to the United States, G.J. Borjas (ed.), pp. 107–21. University of Chicago Press, Chicago. http://www.nber.org/chapters/c0099.pdf.
Lopez, M.H. (2009), “Latinos and Education: Explaining the Attainment Gap.” Pew Hispanic Center Report, pp. 115–16. doi:10.1177/1538192704263548.
Luthra, R.R. and T. Soehl (2015), “From Parent to Child? Transmission of Educational Attainment Within Immigrant Families: Methodological Considerations.” Demography, Vol.52/2, pp. 543–67. doi:10.1007/s13524-015-0376-3.
Morency, J.-D., É.C. Malenfant and S. MacIsaac (2017), Immigration and Diversity: Population Projections for Canada and Its Regions, 2011 to 2036. Statistics Canada. http://www.statcan.gc.ca/pub/91-551-x/91-551-x2017001-eng.htm.
Park, J. and D. Myers (2010), “Intergenerational Mobility in the Post-1965 Immigration Era: Estimates by an Immigrant Generation Cohort Method.” Demography, Vol. 47/2, pp. 369–92. doi:10.1353/dem.0.0105.
Perreira, K.M. and L. Spees (2015), “Foiled Aspirations: The Influence of Unauthorized Status on the Educational Expectations of Latino Immigrant Youth.” Population Research and Policy Review, Vol. 34/5, pp. 641–64. doi:10.1007/s11113-015-9356-y.
Picot, G. and H. Feng (2011a), “Preparing for Success in Canada and the United States: The Determinants of Educational Attainment Among the Children of Immigrants.” Statistics Canada, Analytical Studies Branch Research Paper Series, No. 332, no. 11, pp. 1–28. doi:10.1016/j.aqpro.2013.07.003.
Picot, G. and H. Feng (2011b), “Seeking Success in Canada and the United States: The Determinants of Labour Market Outcomes among the Children of Immigrants.” Statistics Canada, Analytical Studies Branch Research Paper No. 331, Analytical Studies Branch Research Paper Cat. No. 11F0019M No. 331, . Ottawa, ON, pp. 1–29. doi:10.1787/9789264086395-4-en.
Picot, G. and Ar. Sweetman (2005), “The Deteriorating Economic Welfare of Immigrants and Possible Causes: Update 2005.” Statistics Canada: Business and Labour Market Analysis Division. Statistics Canada Analytical Studies Branch Research Paper Cat. No. 11F0019MIE No. 262. Ottawa, ON.
Picot, G., and A. Sweetman (2012), “Making It in Canada: Immigration Outcomes and Policies.” IRPP Study No. 29. Montreal. http://archive.irpp.org/pubs/IRPPstudy/IRPP_Study_no29.pdf.
Postepska, A. (2017), “Ethnic Capital and Intergenerational Transmission of Educational Attainment.” IZA Discussion Paper 10851.
Potochnick, S. (2014), “The Academic Adaptation of Children of Immigrants in New and Established Settlement States: The Role of Family, Schools, and Neighborhoods.” Population Research and Policy Review, Vol. 33/3, pp. 335–64. doi:10.1007/s11113-013-9319-0.
Raleigh, E. and G. Kao, (2010), “Do Immigrant Minority Parents Have More Consistent College Aspirations for Their Children?” Social Science Quarterly, Vol. 91/4, pp. 1083–1102. doi:10.1111/j.1540-6237.2010.00750.x.
Schaafsma, J. and A. Sweetman (2001), “Immigrant Earnings: Age at Immigration Matters.” Canadian Journal of Economics, Research on Immigration and Integration in the Metropolis, Vol. 34/4, pp. 1066–99. doi:10.1111/0008-4085.00113.
Skuterud, M. (2010), “The Visible Minority Earnings Gap across Generations of Canadians.” Canadian Journal of Economics, Vol. 43/3, pp. 862–81. doi:10.1111/j.1540-5982.2010.01598.x.
Smith, J.P. (2003), “Assimilation across the Latino Generations.” American Economic Review, Papers and Proceedings, Vol. 93/2, pp. 315–19.
Smith, J.P. and B. Edmonston (1997), The New Americans: Demographic and Fiscal Effects of Immigration. National Academy Press, Washington, DC.
Statistics Canada (2004), Canada 2001 Census.
Statistics Canada (1962), 1961 Census of Canada.
Statistics Canada (2012) 2011 National Household Survey.
Sweetman, A. and J.C. van Ours (2015), “Immigration: What About the Children and Grandchildren?” In Handbook of the Economics of International Migration, B.R. Chiswick and P.W. Miller (eds.), 1B:1141–94. Elsevier B.V., Amsterdam. doi:10.1016/B978-0-444-53768-3.00021-7.
Taylor, P., D’V, Cohn, J,S. Passel, G, Livingston, W, Wang, C, Funk, A, Gonzalez-Barrera, R, Morin, E, Patten and S, Motel (2013), “Second-Generation Americans: A Portrait of the Adult Children of Immigrants.” PewResearchCenter, Vol. 7, pp. 1–130. www.pewresearch.org.
Trejo, S.J. (2003), “Intergenerational Progress of Mexican-Origin Workers in the U.S. Labor Market.” The Journal of Human Resources, Vol. 38/3, pp. 467–89.
US Census Bureau (1964), Census of Population: 1960, US Department of Commerce.
US Census Bureau (various years), 2007-2011 American Community Survey US Department of Commerce.
US Census Bureau and Bureau of Labor Statistics, Current Population Survey, 1998-2004.
Worswick, C. (2009), “Adaptation and Inequality: Children of Immigrants in Canadian Schools.” Canadian Journal of Economics, Vol. 37/1, pp. 53–77.
Note
← 1. “Immigrants” here refers to adults who migrate to a receiving country with the intention of residing there indefinitely/long term. Caveat: Definitions of “adult” vary, with age-at-arrival minimums from 12 to 25 being common. Researchers not focusing on intergenerational issues – and some who do – do not always establish an age-at-arrival restriction. The entire permanent population is, unless otherwise indicated, divided into four mutually exclusive and exhaustive groups: (adult) immigrants, child immigrants, native born children of at least one immigrant parent (i.e., children of immigrants), and children with two native born parents (i.e., children of non-immigrants). The last category is commonly used as a comparison group. Note that despite the "child" concept, many analyses focus on adults and the classification is not a function of current age. So, for example, a child immigrant may have arrived at age 7 but be age 43 at the time of a survey.