They're not just like the Irish--or the Italians or the Poles, for that matter. The large influx of Hispanic immigrants after 1965 represents a unique assimilation challenge for the United States. Many optimistic observers have assumed--incorrectly, it turns out--that Hispanic immigrants will follow the same economic trajectory European immigrants did in the early part of the last century. Many of those Europeans came to America with no money and few skills, but their status steadily improved. Their children outperformed them, and their children's children were often indistinguishable from the "founding stock." The speed of economic assimilation varied somewhat by ethnic group, but three generations were typically enough to turn "ethnics" into plain old Americans.
This would be the preferred outcome for the tens of millions of Hispanic Americans, who are significantly poorer and less educated on average than native whites. When immigration skeptics question the wisdom of importing so many unskilled people into our nation at one time, the most common response cites the remarkable progress of Europeans a century ago. "People used to say the Irish or the Poles would always be poor, but look at them today!" For Hispanics, we are led to believe, the same thing will happen.
But that claim isn't true. Though about three-quarters of Hispanics living in the U.S. today are either immigrants or the children of immigrants, a significant number have roots here going back many generations. We have several ways to measure their intergenerational progress, and the results leave little room for optimism about their prospects for assimilation.
Before detailing some of those analyses, we should recognize the importance of this question. If we were to discover that, say, Slovenian immigrants did not assimilate over several generations, there would be little cause for alarm. There are simply too few Slovenian Americans to change our society in a meaningful way. Hispanics, on the other hand, have risen from 4 percent to 15 percent of the American population since 1970. The Census Bureau projects that, if there is no change in immigration policy, 30 percent of the nation will be Hispanic by 2050. To avoid developing a large economic underclass, we need to confront the question of whether they will assimilate
The children of Hispanic immigrants (the second generation) actually stay in school much longer and earn a considerably higher wage than their parents. In fact, the Hispanic rate of assimilation from the first to the second generation is only slightly lower than the assimilation rate of more successful groups of immigrants. Most second-generation Hispanics make up nearly as much ground as the children of European immigrants would if they grew up in the same disadvantaged situation.
But the good news ends there, and two problems arise. First, the second generation still does not come close to matching the socioeconomic status of white natives. Even if Hispanics were to keep climbing the ladder each generation, their assimilation would be markedly slower than that of other groups. But even that view is overly optimistic, because of the second, larger problem with Hispanic assimilation: It appears to stall after the second generation. We see little further ladder-climbing from the grandchildren of Hispanic immigrants. They do not rise out of the lower class.
The most straightforward statistical evidence of this stall in Hispanic assimilation comes from the Current Population Survey (CPS), which asks respondents their ethnicity, where they were born, and where their parents were born. From this information we can construct an account of the first generation (foreign-born), the second generation (born in the U.S. with at least one foreign-born parent), and the "3+" generations (born in the U.S. to two U.S.-born parents) among the Hispanic respondents.
This chart shows how Hispanic Americans compare with white natives generation by generation. The annual earnings of second-generation Hispanic men are substantially higher than those of the first generation. However, the 3+ generations have about the same earnings as the second, still well below white natives. No generational progress beyond the second generation is evident.
The educational picture does not look much better. The children of Hispanic immigrants are much better educated than their parents. However, American-born Hispanics still have high dropout and low college-completion rates compared with white natives, and there is little improvement from the second to the 3+ generations. Again, progress stalls.
These results do not depend on the time period considered. Economists Jeffrey Grogger and Stephen Trejo reached the same conclusions when they used CPS data from the mid-1990s for a similar analysis of Mexican Americans. And other datasets tell the same story. One study reported results from the Latino National Political Survey, conducted in 1989 and 1990. Among its striking findings was that the percentage of Mexican-American households with incomes higher than $50,000 rose from 7 percent in the first generation to 11 percent in the second. But the same statistic in the third and fourth generations stayed at 11 percent, at a time when the national rate was 24 percent. Another study, the National Longitudinal Survey of Youth, began tracking a representative sample of young Americans in 1979. By 1993, the Hispanic 3+ generations in that sample had, if anything, slightly worse outcomes than the second generation in terms of wages, educational attainment, and cognitive test scores.
The studies discussed so far are cross-sectional--statistical snapshots captured at single points in time. Since each of the generations being compared lives in the same era, the second-generation respondents are not the actual children of the first generation, nor are the third-generation respondents the children of the second. But longitudinal studies--taking one cohort of Hispanic immigrants, then examining their children and their children's children over several decades--tell a similarly pessimistic story.
Economist James P. Smith pieced together census and CPS data starting in 1940 and ending in 1997. He was able to compare eight different immigrant birth cohorts with their children and grandchildren in later years. Smith found that, contrary to the cross-sectional studies, the Hispanic educational deficit relative to whites did become smaller between the second and 3+ generations. This might indicate an increase in their skills relative to whites, but it might also reflect the trend in the mid-20th century for working-class people to stay in school longer. Did the educational gains for Hispanics affect their relative earning power?
Not by much. The table below, reproduced from Smith's study, shows average Hispanic-American and Mexican-American earnings by birth cohort and generation as a percentage of average white-native earnings. In the six most recent cohorts, the Hispanic panel shows only modest gains from the second to the 3+ generations. For example, Hispanic immigrants born between 1915 and 1919 earned 70.9 percent of what contemporary white natives earned. The children of those immigrants earned 82.3 percent, and the immigrants' grandchildren earned 84.8 percent.
For Mexicans in particular the picture is even worse. In five of the six most recent birth cohorts, the Mexican 3+ generations earn a marginally lower fraction of the white-native wage than does the second generation.
A similar longitudinal analysis was recently conducted by UCLA sociologists Edward E. Telles and Vilma Ortiz. They revived a 1960s-era cross-sectional survey of Mexican Americans by re-interviewing many of the original respondents more than 40 years later. By adding information about the parents and children of the respondents in this second survey wave, the authors were able to construct a longitudinal dataset similar to Smith's. Their results show continued improvement in high-school-graduation rates from the second to the 3+ generations, but small gains in college graduation and stagnant relative wages.
Taken as a whole, the research on Hispanic assimilation presents two possible conclusions. Either Hispanic assimilation will be exceedingly slow--taking at least four or five generations, and probably several more--or it will not happen. In either case, Hispanic immigration will have a serious long-term consequence: The grandchildren of today's Hispanic immigrants will lag far behind the grandchildren of today's white natives.
So why do Hispanics, on average, not assimilate? Theories abound. Popular explanations from the left include the legacy of white racism, labor-market discrimination, housing segregation, and poor educational opportunities. Those on the right tend to cite enforced multiculturalism, ethnic enclaves, and a self-perpetuating culture of poverty. One would need a whole book to sort out these competing explanations, but we can safely say that none of them, even if true, suggests easy solutions. Social scientists have not devised any set of programs that effectively spurs assimilation.
That assimilation has stalled even among third-generation Hispanics growing up today is especially sobering. In the early 20th century, the quality of schools varied greatly, high-school graduation was unusual, travel was relatively difficult, and universities and employers were free to discriminate based on ethnicity. Today all but the worst inner-city schools are well funded, high-school graduation is expected, traveling around the country to look for work is much easier, and affirmative-action programs give preferences to Hispanics. Despite these advantages over earlier immigrants, today's Hispanics have not closed the socioeconomic gap with white natives.
Though continuing research on the barriers to Hispanic assimilation will be valuable, the reality is that no intervention in the foreseeable future will change the very slow and perhaps nonexistent assimilation process into a fast and effective one.
The consequences of a large ethno-cultural group's lagging behind the majority in education and income are significant. In strictly economic terms, perpetually poor immigrants and their descendants will be a major strain on social spending and infrastructure. Health care, public education, welfare payments, the criminal justice system, and programs for affordable housing will all require more tax dollars. When pro-immigration conservatives declare that these government programs should be scaled back or eliminated entirely, I am sympathetic. But a large public sector is a reality that cannot be wished away--we will not be abolishing Medicaid or public schools anytime soon. Immigration policy needs to take that reality into account.
Even if economics were not a concern, the lack of Hispanic assimilation is likely to create ethnic tensions that threaten our cultural core. Human beings are a tribal species, and this makes ethnicity a natural fault line in any society. Intra-European ethnic divisions have been largely overcome through economic assimilation--Irish and Italian immigrants may have looked a bit different from natives, but by the third generation their socioeconomic profiles were similar. Hispanic Americans do not have that benefit.
Persistent ethnic disparities in socioeconomic status add to a sense of "otherness" felt by minorities outside the economic mainstream. Though it is encouraging that Hispanics often profess a belief in the American creed, an undercurrent of this "otherness" is still apparent. For example, a Pew Hispanic Center Survey in 2002 asked American-born Hispanics "which terms they would use first to describe themselves." Less than half (46 percent) said "American," while the majority said they primarily identified either with their ancestral country or as simply Hispanic or Latino. This feeling of otherness probably helps spur explicit ethnic organizing and lobbying. Already there is a long list of Hispanic interest groups--the National Council of La Raza, the Congressional Hispanic Caucus Institute, the U.S. Hispanic Chamber of Commerce, the National Association of Hispanic Journalists, and the Hispanic Lobbyists Association, to name just a handful. If Hispanics fail to assimilate, these groups will remain powerful, and they will continue to encourage Hispanics and other Americans to view our society in terms of inter-ethnic competition. It is difficult to see how a unifying national culture can be preserved and extended in that environment.
Two major changes to our immigration policy are needed to remedy the assimilation problem. First, we should drastically reduce illegal immigration. In the early part of this decade, the illegal-immigrant population saw a net increase of about 515,000 people per year, two-thirds of whom were from Mexico and Central America. The recession appears to have reduced illegal border crossings significantly, but the problem will surely return when our economy improves.
The second change concerns our legal immigration system. While it is important that spouses and minor children of U.S. citizens be allowed to immigrate, our present policy extends well beyond the nuclear family. U.S. citizens can sponsor their parents; their adult children, who may bring their own spouses and children with them; and their adult brothers and sisters, who may also bring their own spouses and children with them. These new green-card holders can then acquire citizenship and bring in their own extended families, perpetuating the cycle. This is "chain migration," and it causes the number of unskilled immigrants in the U.S. to increase swiftly.
It need not continue. Instead of bringing large extended families with limited skills into the U.S., we could specifically select for the qualities--education and work experience, for example--that help immigrants succeed. How would such a system work? We need only look north to see it in practice. Canada assigns points to potential immigrants for various desirable characteristics. For example, holding a graduate degree is worth five times as many points as is holding a high-school diploma. Australia has a similar system. In fact, Canada and Australia take in proportionally three to four times as many immigrants for economic reasons as the U.S. does, and fewer than half as many for family reunification.
Of course, precisely which factors to select for is itself a controversial question, and should be debated in the political arena. But that debate can begin only when we let go of our sepia-toned memories of immigration past. Who gets in really does matter, and we should not let the success of Europeans who came here a hundred years ago obscure that fact.
Jason Richwine is a National Research Initiative Fellow at AEI.