American Institutes for Research/CALDER
CALDER Policy Brief No. 5-0618-1
The United States is in the midst of the second largest wave of immigration in its history. The last thirty years have averaged more than one million documented immigrants per year, changing the demographic composition nationwide. How these immigrants fare, therefore, has significant implications for the nation in terms of, for instance, income and social equality and the long-run competitiveness of the U.S. economy.
In a recent CALDER working paper, David Figlio of Northwestern University and I address this question examining a key predictor of how immigrants are likely to fare as adults: their schooling achievement. Specifically, we analyze the cross-generational differences in educational experiences and outcomes of immigrants in Florida public schools. Florida can be regarded as a mirror for the future demographics of the public school system in the United States because of its position as one of the major destinations for the recent wave of immigration.
We find a general pattern of deteriorating educational outcomes across successive immigrant generations, i.e. first generation immigrants (foreign born students), after a transition period perform better in reading and math tests than do second generation immigrants (U.S. born students to at least one foreign born parent), and second generation immigrants perform better than third generation immigrants (U.S. born students to U.S. born parents). While our analysis does not directly test the validity of different immigrant assimilation theories (as we are unable to look at the differences in educational outcomes between immigrant children and their parents), these findings present evidence that immigrants in this new wave follow a distinct assimilation path than the immigrants who settled in the United States during the Age of Mass Migration between 1850 and 1914.
With this research brief (and the accompanying interactive data visualization), I build off of the CALDER paper, and provide a visual inspection of the trends in white-Hispanic gaps in adult outcomes nationally and in top immigration states (California, Texas, New York, Florida, New Jersey, and Illinois) over the last two decades, with an emphasis on cross-generational differences among Hispanics. In particular, using the Current Population Survey (CPS) March Supplement for years between 1994 and 2016, I examine (1) how the White-Hispanic gaps in educational attainment, employment, income, and health insurance coverage have evolved over the past two decades; and (2) whether these trends vary by the immigrant generation of Hispanics.
I focus on Hispanics because they represent the fastest growing population subgroup in the United States and Hispanic immigration is predicted to remain as one of the major sources of population growth nationwide. Hispanics currently constitute 17 percent of the U.S. population, up from just 9 percent two decades ago. And they are projected to account for one-third of the population by 2050. Therefore, cross-generational differences among Hispanics now will likely have serious ramifications for future generations.
The findings presented in this brief reveal stark differences in adult outcomes between first, second, and third generation Hispanic immigrants (both in levels and long-term trends), and suggest that any analysis of white-Hispanic gaps that fail to identify generational status of Hispanics potentially masks important variation within the Hispanic population. For instance, in 2016, only 12 percent of first generation Hispanic adults had a college degree or higher, compared to 24 percent for second and third generation (Figure 1). Similarly, between 1994 and 2016, the college degree attainment gap between second generation Hispanic adults and whites has virtually remained unchanged, yet the gap between first generation Hispanic adults and whites has jumped from 18 to 30 percentage points. This brief also highlights the need for more detailed parental information in administrative data (e.g., school records) and nationally representative survey data, which would facilitate more disaggregated analysis on Hispanics in the wake of the recent wave of immigration in the United States, and help policy makers better identify populations in need and design better-targeted interventions to improve their outcomes.
It is important to note that my objective in this brief is purely descriptive and the results should not be interpreted as evidence supporting (or against) Hispanic immigrant assimilation. For instance, I cannot rule out the role of differential attrition from Hispanic identity across generations as discussed by Duncan and Trejo (2016) on the observed cross-generational differences because CPS does not report parental racial/ethnic identity. Similarly, the observed trends in cross-generational differences over this time frame could be driven by differences between first generation immigrants who entered the United States in different eras (Massey, 2012).
In the analysis, I use the CPS March Supplement between 1994, when the survey began to collect information on parental place of birth, and 2016, the latest year available. Because I am interested in educational attainment and labor market outcomes, I restrict the sample to individuals between the ages of 25 and 55, most of who had attained their ultimate level of education, yet were not old enough to retire at the time of the survey. Using these data, I identify first, second, and third (or higher) generation immigrants as follows:
I then compare the first, second, and third generation Hispanic immigrants with whites over this time frame along the following dimensions:
In this section, I discuss the findings at the national level whereas the corresponding graphs for top immigrant-receiving states are provided in the interactive graphs.
The share of first generation immigrants among Hispanics has remained relatively flat between 1994 and 2016, slightly increasing from 55 percent in 1994 to 59 percent in the years leading to the Great Recession, and has declined to 53 percent since 2008. In contrast, over this time frame, the share of first generation immigrants among whites has never exceeded 5 percent, whereas the third (or higher) generation immigrants consistently accounted for more than 90 percent of the white population in the United States. This stark difference in the share of recent immigrants among Hispanics and whites highlights the extent of recent immigration from Latin America and the need for more disaggregated analysis when examining the gaps between Hispanics and other racial/ethnic groups.
I then examine the differences between first, second, third (or higher) generation Hispanics and whites along educational attainment, labor market outcomes, and health insurance coverage, and how these gaps have evolved over the last two decades. The overarching conclusion of this analysis is that first generation Hispanics significantly lag second and third generation Hispanics along all these dimensions, with late-arriving first generation immigrants (those who arrived in the United States after the age of 10) faring the worst whereas those who arrive early in life achieving similar outcomes as their second and third generation counterparts.
Educational attainment: Figure 1 presents the cross-generational differences in educational attainment (% with less than high school diploma on the left, and % with college degree or more on the right) among Hispanics, with the bold lines representing the averages for all Hispanics and whites nationally. Looking at the overall gaps between whites and Hispanics, Hispanics seem to have gained ground compared to whites over this time frame in terms of the % of individuals with less than high school degree. In particular, in 1994, 38 percent of all Hispanics in the United States had less than high school diploma compared to 8 percent among whites. Since then, this gap has declined to 21 percentage points (25 percent among Hispanics versus 4 percent among whites).
In contrast, the white-Hispanic gap in the right-tail of the educational attainment distribution has increased over this time frame. In 1994, 10 percent of all Hispanics nationwide had a college degree (2-year or 4-year) or higher compared to 28 percent for whites. In 2016, this gap increased to 25 percentage points (42 percent for whites versus 17 percent for Hispanics).
These figures also reveal significant variation in educational attainment within the Hispanic population by immigrant generation. For instance, first generation Hispanic immigrants have significantly lower levels of education compared to second and third generation. In 1994, half of first generation Hispanics in the United States had less than a high school degree and only 10 percent had a college degree or higher.
While educational attainment has improved among first generation Hispanics since then (% less than high school degree down to 38 percent in 2016), the attainment gaps between the first generation and higher generation Hispanic immigrants have persisted, and even widened on the higher end of the spectrum. Between 1994 and 2016, the share of second generation Hispanics with a college degree or higher doubled from 12 percent to 24 percent (10 percent to 23 percent among third generation or higher immigrants) in stark contrast to the small increase for first generation Hispanics from 10 percent to 12 percent. This pattern also indicates that the widening gap in college completion between whites and Hispanics is entirely driven by the widening gap between first generation Hispanics and whites.
Employment and Income: A possible implication of these observed gaps in educational attainment across Hispanic immigrant generations is cross-generational differences in labor market prospects. Figure 2 presents the white-Hispanic gaps in labor market outcomes along with the cross-generational differences among Hispanics between 1994 and 2016. In 1994, the employment rate among Hispanics was 70 percent, which increased to 76 percent right before the Great Recession. The Great Recession took a toll on employment across-the-board, but Hispanic employment took a slightly larger hit than white employment for all immigrant generations. Overall, the white-Hispanic gap in employment rate has declined from 11 percentage points in 1994 to 5 percentage points in 2016. Among the Hispanic population, first and second generation immigrants had significantly lower employment rates compared to the third generation in 1994 (68 and 67 percent respectively versus 78 percent), yet the second generation has surpassed the third generation over this time frame. Further inspection reveals that these gaps were mainly due to differences in labor market participation rates rather than differences in unemployment rates.
Income gaps between whites and Hispanics have slightly declined between 1994 and 2016 (from roughly $16,200 in 1994 to $15,000 in 2016). The right panel of Figure 3 also reveals that this decline in the income gap was mainly driven by the increase among the second generation Hispanics over this time frame.
Health Insurance Coverage: In the midst of the heated debates surrounding the effects of the Affordable Care Act (ACA) and the recent evidence in the literature showing long-term benefits of health insurance coverage early in life (Miller and Wherry, forthcoming), another interesting question is how health insurance coverage for first, second, and third generation Hispanics have changed since the enactment of this legislation. Figure 3 presents the cross-generational differences in health insurance coverage among Hispanics, and illustrates the dramatic change in white-Hispanic gaps in health insurance coverage following the enactment of ACA. Before ACA, coverage rate among Hispanics ranged between 60 to 67 percent – first generation immigrants had the lowest coverage rate (46-58 percent) perhaps due to lower employment rates and lower eligibility rates for Medicaid and Medicare whereas the second and third generation had significantly higher rates (75-81 percent). Since 2013, health insurance coverage rate among Hispanics increased roughly by 16 percentage points (or 26 percent), with the first generation Hispanics experiencing the largest improvement (20 percentage points or 41 percent). Before the enactment of ACA in 2010, Hispanics were 26 percentage points less likely to have health insurance coverage compared to whites, yet this gap declined to 13 percentage points in 2016.
In what follows, I break down the analysis further by examining the cross-generational differences among Hispanics along three dimensions of interest: (1) gender; (2) age at arrival for first generation immigrants; and (3) country of origin for first and second generation immigrants.
Males versus Females: Gender gaps in education and labor market outcomes have been well documented in the literature. The question in this context then becomes whether these gaps vary by immigrant generation among Hispanics. Gender breakdown in white-Hispanic gaps in these outcomes presented in Figure 4 reveal some interesting differences. For instance, consider the gaps in the share of individuals with college (or higher) degree between males and females. Among males, the white-Hispanic attainment gap has increased modestly from 20 percentage points in 1994 to 23 percentage points in 2016. On the other hand, this increase was much more dramatic among females, jumping from 16 percentage points to 25 between 1994 and 2016.
Over this time frame, both white and Hispanic females surpassed their male counterparts in college degree (or higher) attainment. For instance, in 1994, 26 percent of all white females had attained a college degree or higher compared to 31 percent among males, yet in 2016, these numbers were 44 percent for females versus 39 percent among males. Patterns are similar for Hispanics but the increase among females is considerably smaller – female attainment increased from 10 percent to 19 percent over this time frame compared to 11 percent to 16 among Hispanic males. For both female and male Hispanics, the improvement in attainment is much more muted among first generation immigrants compared to second and third generation.
Another interesting pattern in this context is the breakdown of the white-Hispanic differences in labor market outcomes by gender. Figure 4 suggests that the observed white-Hispanic gaps in employment is mainly driven by differences among females. In 1994, the employment rate among Hispanic males was 83 percent compared to 88 percent for white males, and this gap virtually vanished by 2016. Among females, on the other hand, the white-Hispanic employment gap declined from 17 percentage points in 1994 to 12 in 2016, yet remained sizable.
These figures also reveal significant gender differences among first generation Hispanic immigrants. While foreign born Hispanic males had comparable employment rates to those of whites, the female employment rate among Hispanic first generation immigrants was 18 percentage points lower in 2016 (down from 25 percentage points) than female whites. This gap was 5 and 6 percentage points for third and second generation Hispanic females in 2016, indicating that observed white-Hispanic gap in Figure 3 was mainly due to the gap between white females and first generation Hispanic female immigrants.
Despite similar employment rates, white-Hispanic gap in income among males is much larger than the gap for females. In 2016, the median Hispanic male earned $31,000 compared to $50,000 for white males (this gap was roughly $22,000 in 1994) whereas the median Hispanic female earned $17,000 versus $30,000 for the median white female (compared to a gap of $11,500 in 1994). Once again, these gaps are much larger for first generation Hispanic immigrants.
Childhood Arrivals versus Adult Arrivals: An overarching conclusion of the extant literature that examines the differences between recent immigrants and natives is that there is significant heterogeneity in the outcomes and experiences of recent immigrants by the age at arrival into the United States (e.g., Figlio and Ozek, 2016; Schwartz and Stiefel, 2006; and Conger et al., 2007). To inspect how differences between first and second generation Hispanic immigrants change based on the age at arrival for first generation, I break down the first generation into 3 groups: (1) those who arrived before the age of 10; (2) those who arrived between ages 10 and 18; and (3) those who arrived after the age of 18. In 2016, 14, 29, and 57 percent of all Hispanic first generation immigrants aged between 25 and 55 fell into the first, second, and third category respectively.
Figure 5 reveal significant differences among the first generation Hispanics based on age at arrival, with those who arrived early in childhood achieving better outcomes compared to teen arrivals and adult arrivals. For instance, in 2016, 23 percent of childhood arrivals had less than a high school degree whereas the same numbers for teen arrivals and adult arrivals were 43 and 40 percent respectively. Childhood arrivals also had slightly higher rates of college degree (or higher) attainment (15 percent) than adult arrivals (13 percent) whereas the teen arrivals had the lowest attainment rate (8 percent).
Another interesting finding is that the gaps in educational attainment between childhood arrivals and teen and adult arrivals have closed between 1994 and 2016. In 1994, 24 percent of childhood arrivals had less than a high school degree (versus 50 percent for teen arrivals and 56 percent for adult arrivals), and 18 percent of childhood arrivals had a college degree or higher (versus 7 percent for teen arrivals and 10 percent for adult arrivals). Childhood arrivals also had better educational attainment than second generation Hispanics two decades ago, yet they have been surpassed by the second generation since then.
Mexican versus Non-Mexican Origin: Another common finding in the prior literature in this context is that the outcomes of recent immigrants vary considerably by country of origin (e.g., Figlio and Ozek, 2016; Schwartz and Stiefel, 2006; and Conger et al., 2007). While I can’t conduct a detailed country-by-country analysis due to sample size issues in CPS, I break down the analysis for first and second generation Hispanics by Mexican origin in Figure 6. For first generation immigrants, Mexican origin implies that the individual was born in Mexico whereas for the second generation, it implies that at least one parent was born in Mexico. In 2016, 62 percent of all first generation Hispanics nationwide were born in Mexico, and 55 percent of all second generation Hispanic immigrants had at least one parent born in Mexico.
The main conclusion of this analysis is that the differences between first and second generation are significantly larger for Hispanic immigrants with Mexican origins. For example, in 2016, 45 percent of all first generation immigrants from Mexico had less than a high school degree compared to 27 percent for first generation Hispanic immigrants from other countries. In contrast, these numbers are almost identical for second generation Hispanic immigrants with Mexican origins and those from other countries.
Cross-generational differences in college degree (or higher) attainment between first and second generation Hispanics from Mexico are also larger compared to immigrants from other countries. Between 1994 and 2016, this difference between first and second generation has increased for both groups, but remains larger for immigrants from Mexico. In particular, in 2016, second generation Mexican immigrants were 13 percentage points more likely to have attained a college degree (or higher) than their first generation counterparts whereas this gap is 8 percentage points for non-Mexicans.
Immigrants and children of immigrants currently account for nearly a quarter of all school-aged children in the United States, and are projected to account for one-third by 2050 (Passel, 2011). However, as noted by Suárez-Orozco et al. (2015) in a recent WT Grant Foundation Inequality Paper, immigrant youth have long been at the margins in the design and implementation of policies and programs and research to inform them.
The lack of research in this context is mainly driven by data limitations. In particular, data systems that contain detailed information about educational experiences (e.g., state administrative data systems) typically fail to identify immigrants, whereas nationally representative surveys that contain information about immigrants (e.g., National Education Longitudinal Survey) fail to incorporate educational outcomes beyond attainment.
For our recent work in Florida, we overcame these data limitations by linking school records with birth records, which enabled us to identify first, second, and higher generation immigrants in our analysis. Using these unique data, we have addressed a number of research questions on immigrant students including an analysis of cross-generational differences, the effects of Haitian earthquake refugees on incumbent students, and the role of cultural values in explaining differences in educational outcomes among immigrants. These studies illustrate the potential of administrative records in immigrant research, and open the door to rigorous empirical analysis of the causal effects of some of the major educational policy initiatives of the last two decades on immigrant-origin youth such as English learner programs and test-based accountability policies that carry significant ramifications for low-performing students (such as grade retention) and disproportionately affect recent immigrants.
Suggested Citation: Özek, Umut (2018). A Closer Look at the Trends in White-Hispanic Gaps in Education and Beyond: 1994-2016. CALDER Research Brief.
Conger, Dylan, Amy E. Schwartz and Leanna Stiefel (2007). Immigrant and Native-born Differences in School Stability and Special Education: Evidence from New York City. International Migration Review 41(2), 402-431.
Duncan, Brian and Stephen Trejo (2016). The Complexity of Immigrant Generations: Implications for Assessing the Socioeconomic Integration of Hispanics and Asians. NBER working paper 21982.
Massey, Douglas S. (2012). Immigration and the Great Recession. Stanford, CA: Stanford Center on Poverty and Inequality.
Miller, Sarah and Laura R. Wherry (forthcoming). The Long-Term Effects of Early Life Medicaid Coverage. Journal of Human Resources.
Schwartz, Amy Ellen and Leanna Stiefel (2006). Is There a Nativity Gap? New Evidence on the Academic Performance of Immigrant Students. Education Finance and Policy 1 (1), 17–49.
Suárez-Orozco, Carola, Hirokazu Yoshikawa, and Vivian Tseng (2015). Intersecting Inequalities: Research to Reduce Inequality for Immigrant-Origin Children and Youth. A William T. Grant Foundation Inequality Paper.
Özek, Umut and David Figlio (2016). Cross-Generational Differences in Educational Outcomes in the Second Great Wave of Immigration. CALDER Working Paper: 162.