To the Editors:

Although Christopher Jencks’s recent two-part essay [“Who Should Get In?,” NYR, November 29 and December 20, 2001] exhibits the intellectual dexterity that readers have come to expect from him, his analysis has overall an unjustifiably gloomy tenor concerning the integration prospects of contemporary immigrants and their children.

He dismisses too facilely the possible continuities between the successes of the immigration past, symbolized by the eventual assimilation of Southern and Eastern European groups, and what appears as the problem-pocked landscape of the present. Yet only in the last few decades has the assimilation of the descendants of pre– World War I immigrants been adequately recognized—the blinkered view that Italians, for example, were stuck in the working class (to say nothing of crime) persisted well after substantial upward mobility had begun. Perhaps the view of the contemporary scene is equally clouded.

In decoupling past and present, Jencks relies excessively on economic data and, in particular, on a comparison of immigrant earnings (in relation to those of natives) in 1910 and 1998, which comes from the work of the Harvard economist George Borjas. For a variety of reasons, this comparison is less than definitive—for one thing, the earnings of workers in 1910 were not directly measured (by the US Census) and must be inferred indirectly by a procedure that may distort the economic situations of recently arrived immigrants.

In any event, it is not clear why one should privilege exclusively the economic dimension. The demographic one, which seems just as relevant, especially given the ethnic and religious divisions of the earlier period, yields a comparison pointing in the other direction. That is, in relation to the receiving population, the immigration of the 1830–1925 period was much larger—several times, including the 1900–1909 period, the number of immigrants arriving in a single decade approached 10 percent of the population. There is no decade during the current immigration era that has come close to that figure. In this respect, the challenge of absorbing immigrants was greater in the earlier period than today.

Certainly, the incorporation of contemporary immigrants and the second generation will not simply replicate that of the earlier era, and much about it, even in the near future, remains unpredictable. But Jencks’s emphasis on downward assimilation seems strangely one-sided. To be sure, it is often claimed as a distinguishing feature of the current era. Yet there were surely parallels, including some not-so-benign effects of Americanization, in the earlier period as well. For the contemporary and second generation, there are no convincing data to demonstrate that socioeconomic stagnation characterizes it en masse. Reynolds Farley and I are publishing a paper in the International Migration Review that is based on an analysis of 1998– 2000 Current Population Survey data. The analysis shows unmistakable socioeconomic advance in the second generation, especially for groups, such as the Mexicans, that enter with low levels of human capital. Perhaps an analysis at the end of the prosperous 1990s may overstate the upward movement to be expected on average in the future. But such a concession is hardly proof of wholesale downward assimilation.

Shedding light on the positive aspects of contemporary immigration, in part on the basis of a nuanced understanding of the immigration past, is as important as calling attention to the social problems Jencks details. A more balanced view may help us to see the breadth of immigrants’ contributions to American society (and vice versa) and to recognize that many are not strictly economic; immigration has, in the past and present, expanded our sense of who we are and fueled a cultural and social dynamism that is very much a part of the American condition.

Richard Alba

Albany, New York

To the Editors:

Christopher Jencks’s first essay on immigration, “Who Should Get In?” [NYR, November 29, 2001] has many virtues, including erudition, range, and grace, but also a subtle flaw. This lies in the unexamined premise that the immigrants themselves are the causal agents. According to this premise, illiterate, low-skilled, and otherwise disadvantaged persons first, somehow, arrive. They then bid down the wage rates for less-skilled occupations. And, if successful, they displace native workers who had previously held those jobs but who reject them once wages fall.

Quantity is cause, price is effect. This is the famous mechanism of supply and demand, which economists love. It leads directly to a policy focus on permissible quantities of unskilled labor—“Who Should Get In?”

But instead suppose that the employers, rather than immigrants, are the active agents. Suppose their action is to cut wages—outsourcing, or an anti-union campaign, say, or perhaps a legislative maneuver that blocks a rise in the minimum wage and so causes its real value to fall. Now for native workers the worst jobs no longer provide adequate pay. Immigrants may then be recruited, directly or through the grapevine, to fill them. They will flood the regions (Los Angeles, Texas) where sweatshops flourish, while remaining scarce in union towns like Detroit. And they will come from low-wage countries, like Mexico.


This sequence has very different implications for policy. For, if only the wage structure had been maintained in the first place—through strong unions, robust minimums, and other measures—then the excess demand for immigrants and the phenomenon of displacement would not occur. It is no coincidence that in 1970 there was no immigrant–native wage differential and yet no migrant wave. The real value of the federal minimum hourly wage, back then, exceeded seven dollars in today’s values, and a much higher fraction of the workforce belonged to unions.

Harvard University itself provides an example of wage cuts followed by a changing workforce. In recent years Harvard has driven down the real wage of its staff; today mean custodial pay there is $9.55 per hour, compared to $11 in 1994. This is not because immigrants have forced down “market” wages in Cambridge; just two subway stops away, MIT pays a starting custodial wage of $14.39 per hour, according to the Boston press. Meanwhile, though, the Harvard custodial staff has indeed become much more Hispanic in recent years as wages fell; two thirds of the janitors there today are immigrants.

In short, faced with such employers, politically weak groups accept jobs on terms that, in other times, American labor would not stand for, nor American opinion tolerate. And accordingly, the solution to our immigration dilemmas may be closer to home than we think.

James K. Galbraith


Lyndon B. Johnson School of Public Affairs

University of Texas at Austin

Austin, Texas

To the Editors:

As Christopher Jencks suggests [“Who Should Get In? Part II,” NYR, December 20, 2001], repeating the spirit of the 1986 Immigration Reform and Control Act, with its proven compassion and its intended firmness, may improve the lot of immigrants and reduce their burden to American society. We have already learned, however, that labor market forces will perpet- uate illegal immigration even in the face of the most draconian employer restrictions and that many of the immigrants’ children will continue to face challenges in the US. How then can we act on the socioeconomic forces driving immigration and on the condition of the “second generation?”

One answer may lie in the institutions through which immigrant families interact with their countries of origin. Jencks identifies the close ties between Mexicans and their home villages—obstacles to their as- similation and to the prospects for their children. In fact, many Mexican immigrants—and their Latin American neighbors—form village and regional-based associations that join hands to build schools, hospitals, and businesses in their homelands. Through these efforts, the immigrants work to improve the conditions that have forced so many of them to migrate. At the same time, they form soccer leagues and cultural groups that tie the second generation to its parents’ culture and offer it alternatives to gangs and other social ills. Finally and most importantly, these associations build leaders within both generations, who will become more productive members of whichever society they eventually settle in.

The European Union has embraced such associations among its immigrant groups and begun covering up to 80 percent of immigrant-led development projects in North Africa. Likewise, the United States government—and charitable foundations—could channel funds and technical assistance through immigrant institutions here. Doing so would act more directly on the deeper problems that quotas, visas, and security systems can’t reach.

Michael Kerlin

Economic Development

Project Manager

Nuestra Comunidad

Development Corporation

Boston, Massachusetts

Christopher Jencks replies:

In light of the passions that immigration usually arouses, all the letters about my articles on immigration were unusually thoughtful and constructive.

Richard Alba is certainly right to say that one can tell both optimistic and pessimistic stories about immigrants. Both stories are correct. As he says—and as I myself said—most second-generation immigrants are getting more schooling and earning more money than their parents. But second-generation immigrants are also doing worse than their parents in some other respects, like keeping their families together and keeping their children alive. We need to keep both stories in mind rather than asking which one is correct.

Alba is also right that the assimilation of pre-1914 immigrants was not generally recognized until the 1970s. That was partly because “old stock” Americans had negative stereotypes about these immigrants, but it was mainly because assimilation really took three generations. Those who advocate admitting more immigrants today need to be more candid about how long assimilation takes. Many enthusiasts dismiss this issue on the grounds that “assimilation” is not even a legitimate goal, but without assimilation there is no obvious reason why Americans should want to share “their” country with others.

When I first looked at George Borjas’s estimates of the economic gap between immigrants and natives in 1910 I shared Alba’s doubts about their validity. The Census Bureau did not ask Americans about their income until 1940, so Borjas had to infer different groups’ income by combining Census data on their occupations with other surveys that collected data on wage differences between occupations. To do this Borjas had to assume that immigrants and natives who worked in the same occupation had the same average wage.


Joel Perlmann, a historian at Bard College’s Levy Institute who was also skeptical about my earnings estimates for 1910, has now used the 1940 Census to compare the 1939 earnings of forty-five-to-sixty-four-year-old men born in Italy, Poland, and other Central European countries to the earnings of native whites. He finds that the immigrants earned almost as much as natives in the same occupations in 1939. These men were between fifteen and thirty-four at the time of the 1910 Census. Since wage gaps usually grow with age, Perlmann’s findings suggest to me that the wage gap between immigrants and natives in the same occupation was also small in 1910.

But Perlmann’s findings, which are available from the Levy Institute in Working Paper 343, also raise a question about the accuracy of the earnings data for occupations in the early 1900s. Consider the Italians, who were the least-skilled major group of European immigrants in 1910. Borjas calculated that Italians earned 83 percent of what native whites earned. The overlap between immigrant and native occupations did not change much between 1910 and 1940, at least among men old enough to have been in the labor force in 1910. But the earnings gap between immigrant and native occupations widened from 17 to 26 percent. The same pattern recurs for immigrants from Eastern Europe. My reading of Perlmann’s findings is that the earnings gap widened because of high unemployment in less-skilled occupations during 1939. A decade later, when unemployment was much lower, the gap between these occupations looked much like the estimated gap in 1910.

Perlmann’s findings seem to me consistent with the argument that Southern and Eastern European immigrants were far less economically disadvantaged in 1910 than today’s Mexican immigrants. But his findings also suggest that the Depression hit Southern and Eastern European immigrants harder than natives. In the long run, however, the Depression may have hastened immigrants’ economic assimilation. The Depression led to unionization in heavy industry, which raised second-generation immigrants’ wages after 1945. In addition, the dearth of jobs during the 1930s kept many second-generation immigrants in school, allowing them to move into better jobs when good times returned.

Maybe I am missing something, but I have never seen the relevance of the fact that immigrants constituted a somewhat larger fraction of the US population in 1910 than in 2000. “Immigrants” seem to me too diverse to constitute a useful analytic category for thinking about assimilation. If you worry about linguistic assimilation, for example, it is not helpful to know how many Americans speak a language other than English at home. If immigrants are very numerous but they all speak different languages, they will all learn English because they will not be able to get by in their native language. If millions of immigrants all speak the same language, they may not need English. So if you want to know whether the fact that Italians almost all learned English tells you anything about the likelihood that second-generation Mexicans will learn English, you need to know whether the Italian-speaking minority was as large in 1910 as the Spanish-speaking minority is today. The answer to that question is no.

James Galbraith argues that employers actively recruit immigrant workers, and I certainly agree. Many employers also campaign for more open borders, because they think immigrants make better workers than the natives they can hire at the same wage. But while I favor forcing employers to provide better wages and working conditions, I do not see why this would induce employers to hire natives instead of immigrants. Surely raising wages for unskilled work would lure even more immigrants to the United States, just as it has lured them to Western Europe. And if employers prefer unskilled immigrants to the unskilled natives who currently apply for $6-an-hour jobs, why would these employers’ preferences change when they had to pay $9 an hour?

It is true, as Galbraith says, that MIT now hires natives for its $14.39 custodial jobs, while Harvard hires immigrants for its $9.55 custodial jobs. But $14.39 is more than twice the current minimum wage even in Massachusetts. If the minimum wage for all workers were $14.39, firms that offered “only” $14.39 would still get the least-desirable applicants. Firms that wanted better workers would have to offer $16 or $18 an hour. Faced with a choice between America’s least-desirable natives and unskilled immigrants, employers who paid the minimum, whatever it happened to be, would still prefer the immigrants. If we don’t want employers to make this choice, we have to curtail the flow of unskilled immigrant workers.

Michael Kerlin argues that employers will keep hiring illegals even in the face of “draconian employer restrictions” of the kind that I advocated. He may be right. But the United States has never implemented a system of employer sanctions that was at all draconian, so we cannot know whether such a system would work until we try it.

I do agree with Kerlin that promoting third-world development, especially in Mexico and Central America, is crucial if we want to slow immigration. But development requires a political and legal regime that encourages long-term investment, limits corruption, and adopts sensible policies. I am not sure how channeling US aid through emigrant organizations will increase Latin America’s chances of pursuing these goals, much less achieving them.

I am also sorry that Kerlin read my article as arguing that maintaining ties to a Mexican village discouraged Mexicans living in the United States from assimilating. My intention was to suggest that the causal arrow ran the other way. Poor Mexicans maintain ties to their home villages because they want to see themselves as relatively successful Mexicans rather than as relatively unsuccessful Americans. If we want immigrants to assimilate quickly, we need to admit people who will do well here.

This Issue

May 23, 2002