The Wall Street Journal published an editorial the other day arguing that new immigrants to the U.S. are "assimilating" just as well as previous cohorts of newcomers, writing that "today's immigrants are acculturating and moving up the economic ladder like previous generations."

It's certainly true that the children and grandchildren of immigrants are much better off than their parents and grandparents. However, new research from Harvard economist George Borjas suggests that the living standards of first-generation immigrants are rising toward the levels of native-born Americans more slowly than in previous decades.

While moving to the U.S. from a poor country is often a good way to raise one's income, adults who migrate here are usually paid much less than those born here. Historically, this wage gap (among adults of the same age) shrank as immigrants spent more time living in the U.S. According to Borjas, there was a significant change sometime in the 1980s, as illustrated by the accompanying chart.

After presenting this new empirical finding, Borjas looks for causes. One obvious possibility is that globalization and technological change, which we know play at least some role in increasing the inequality of income distribution, disproportionately affects adult migrants. While it turns out that "changes in the wage structure can account for about a third of the observed decline in entry earnings" for new immigrants, "the finding of a decline in the rate of economic assimilation is impervious to the use of skill-specific price deflators."

Change in average wages of adult immigrants relative to average wages of native-born workers. Source: George Borjas
Change in average wages of adult immigrants relative to average wages of native-born workers. Source: George Borjas

Borjas also looked at the relative earnings of immigrants and natives within occupational categories, as well as immigrants and natives who started out in the same percentile of the income distribution. He concludes that, "carefully accounting for the nature of period effects cannot explain why the rate of economic assimilation declined so dramatically for the immigrant cohorts that entered the United States after the mid-1980s." That leaves us searching for other possibilities.

One is geographic variation. Wages vary widely across the U.S. Per-capita personal income in Connecticut is almost $58,000 but only about $33,500 in Mississippi. So if the geographic distribution of immigrants radically changed in the 1980s, it could theoretically produce Borjas's empirical results. In practice, however, Borjas finds that the changing distribution of recent immigrants within the U.S. "barely changes" his earlier results.

Perhaps it matters where the immigrants came from. Who knows, maybe people from certain countries are destined to earn more or less money than people from other countries. Yet Borjas found that the rate of wage catch-up slowed down "even within national origin groups." So even though immigrants from China who arrived in the late 1990s did much better than immigrants from Mexico, those Chinese immigrants experienced slower wage gains than their predecessors. As Borjas says, "changes in the national origin mix of immigrants cannot explain why the more recent cohorts have a much lower rate of economic assimilation."

Borjas is left with something else: Recent immigrants have accumulated relatively less "human capital" than earlier generations of immigrants. In particular, they are less likely to learn English than those who came before them. This is all the more remarkable considering that the share of newcomers who arrive in the U.S. already knowing English (about one-third) hasn't really changed since the 1970s. The second accompanying chart, from Borjas's paper, shows the change in the share of immigrants who know English. As you can see, there is a marked difference between those who arrived before 1980 and those who arrived afterward, mostly because of differences in the first 10 years:

Increase in English-language proficiency among different immigrant cohorts. Source: George Borjas
Increase in English-language proficiency among different immigrant cohorts. Source: George Borjas

Why could this have happened? Borjas has one explanation that stands up to the facts, although it is incomplete: Large numbers of earlier immigrants from the same countries has reduced the need for new immigrants to learn English. "Immigrants who enter the country and find a large welcoming ethnic enclave have much less incentive to engage in these types of investments since they will find a large market for their pre-existing skills," Borjas writes.

The data seem to partially support this hypothesis. Borjas finds that "the increasing size of the immigrant national origin group accounts for about 20 percent of the decline in the rate of economic assimilation" and "explains about 40 percent of the decline in the rate of (language) human capital investments across the cohorts." Intriguingly, this result holds true even when Borjas excludes Hispanic immigrants from his study.

Reading this paper, I kept wondering what the results would have looked like if Borjas had been able to study comprehensive data on immigration, incomes, educational attainment and English-language proficiency going back two centuries, rather than a mere four decades. Perhaps the same patterns could have been observed in the late 1800s and early 1900s, when my ancestors came here. Unfortunately, no such data exists.

Either way, Borjas's findings are important. Immigrants do better when they learn English, yet the incentives for them to do so decline as more arrive. If we plan on letting in more immigrants, as most economists recommend, we may need to implement policies that push new Americans to learn English to offset this troubling trend.

(Matthew C. Klein is a contributor to the Ticker. Follow him on Twitter.)