Posted on July 14, 2015

Are Low-Skill Immigrants Upwardly Mobile?

Jason Richwine, Real Clear Policy, July 8, 2015


Many people take for granted that today’s low-skill immigrants will be just like the Irish and Italians of Ellis Island lore, coming to the U.S. as menial laborers but rising to the middle class within a few generations. But it is a serious mistake to assume that today’s immigrants will assimilate as quickly as others did in the past. In fact, the data from recent immigration waves tell a pessimistic story about upward mobility.

Let’s focus on immigrants from Latin America, since they form the bulk of less-educated immigrants from the past 50 years and are easier to track across generations. The most straightforward evidence on Hispanic assimilation comes from cross-sectional data sets produced by the Census Bureau. All of these data sets reveal that economic and educational progress for Hispanics stalls after the second generation, leaving a substantial gap between white Americans and Hispanic Americans.

Take education. According to the Current Population Survey, about 37 percent of white adults have a four-year college degree, while just 12 percent of Hispanic immigrants do. In the Hispanic second generation (meaning U.S.-born with immigrant parents), the rate of college graduation goes up to 23 percent, but it is just 18 percent in the Hispanic “third-plus” generation (meaning U.S.-born with two U.S.-born parents). The data on income, poverty, and household welfare receipt show a similar generational pattern.

In addition, the norms of hard work and discipline exhibited by Hispanic immigrants appear to slip in subsequent generations. This is sometimes referred to as “downward assimilation” to the American underclass. For example, just 10 percent of Hispanic immigrant men are out of the labor force, but that number goes up to 21 percent in the third-plus generation. As for crime, Hispanic immigrant men are 27 percent more likely than white men to be listed in the American Community Survey as institutionalized (usually meaning in prison), while U.S-born Hispanic men are fully 157 percent more likely.

How might this picture be too pessimistic? One possibility is that cross-sections are misleading, since each of the generations being compared lives in the same era. Maybe we would see more assimilation if we go back in time to an original immigrant cohort, then follow the actual children and grandchildren of those immigrants.

The good news is that this longitudinal approach does show gains in education for Hispanics between the second and third generations. The bad news is that relative wages remain basically stagnant, just as in the cross-sectional analysis. The further bad news is that even the educational improvement may be a mirage–does it reflect real skill gains relative to whites, or just the general mid-20th century trend for working-class people to stay in school longer?


If we are serious about improving social mobility in the U.S., then importing millions of low-skill immigrants (of any nationality) with uncertain prospects for advancement cannot be the right policy. No one wants the U.S. to be defined by the gated communities and rigid class lines common in Latin America–especially when the class lines are tinged by skin color–but that is the outcome we risk.