The Top Ten Symptoms of Immigration

By George Borjas on November 1, 1999

This Backgrounder is adapted from Dr. Borjas’ latest book, Heaven’s Door: Immigration Policy and the American Economy, which presents a comprehensive, accessible, and up-to-date account of the economic impact of recent immigration on America. Dr. Borjas reveals that the benefits of immigration have been greatly exaggerated and that, if we allow immigration to continue unabated and unmodified, we are supporting an astonishing transfer of wealth from the poorest people in the country, who are disproportionately minorities, to the richest.

As the 21st century begins, the United States is about to embark yet again upon a historic debate about the type of immigration policy that the country should pursue. And, as always, the debate will inevitably focus on two questions: How many people should the United States admit? And, since there are many more persons who will want to migrate to the United States than the country is willing to admit, which of the visa applicants should the country accept?

Consider some of the issues at the core of today's immigration debate. There is a great deal of worry that immigrants make extensive use of social services and don’t "pay their way" in the welfare state. There is also anxiety that labor market competition with immigrants might have a harmful impact on the economic well being of some American workers. And there is the traditional concern over assimilation: Will today’s immigrants find it harder to assimilate than earlier waves?

On the benefit side, it is sometimes argued that immigration spurs economic growth, perhaps by contributing to the creation of new industries, such as in Silicon Valley. The large-scale migration of less-skilled workers may also reduce the prices that American consumers pay for many goods and services. Moreover, immigrants may do jobs that natives refuse to take, so that certain industries in the country — such as California’s agriculture industry — would likely disappear if immigrant labor were not available.

These are all valid issues that should be part of the immigration debate. Unfortunately, they are also the symptoms of pursuing particular types of immigration policies. By arguing over the validity of these symptoms — do immigrants use a lot of welfare; are consumer prices lowered by immigration — the immigration debate is, in a sense, worrying about the height of the trees in the forest, rather than about the shape of the forest itself.

Heaven’s Door attempts to shift the terms of the immigration debate — away from arguing endlessly over the validity of the particular symptoms and toward the fundamental questions: How many immigrants does the United States want? And which types of immigrants should the country admit? By framing the debate in this fashion, I hope to clarify what is really at stake which is nothing less than our conception of what the United States is about. It is futile to think about how many and which immigrants to admit unless one first has some objectives in mind. What is it that the American people want immigration to do for the country?

The Top Ten Symptoms

Participants in the debate over immigration policy typically use an array of statistics, many of them drawn from the latest research by economists, as weapons in this debate. It is instructive to list the top ten symptoms (not necessarily in order of importance) that frame the immigration debate, and that will likely determine its direction.

1. The number of immigrants entering the United States is at record levels.

Although the United States has admitted immigrants throughout its entire history, the number of immigrants admitted into the country has fluctuated greatly over time. Eras of large migration, for instance, were followed by decades of rest, during which time the immigrant waves were presumably assimilated and incorporated into the American mainstream.

Surprisingly, relatively few immigrants (only about 10 million) entered the country between 1820 and 1880. The huge flow that has come to be known as the Great Migration began around 1880 and continued until 1924, bringing with it about 26 million immigrants. The immigration restrictions imposed in 1924, as well as the Great Depression, reduced the immigrant flow to a trickle by the 1930s. Since then, the number of immigrants has increased steadily, with the increase accelerating in the 1970s and 1980s. By the late 1990s, nearly one million persons entered the country legally each year and another 300,000 entered the country illegally. (See Figure 1.)

Because of the increasing number of immigrants and the lower fertility rate of American women, immigration in the 1990s accounted for at least a third of population growth. And this large demographic impact clearly justifies calling the large immigrant wave that began to enter the United States after 1965 the Second Great Migration.

Graph: Legal Immigration to the United States by Decade, 1810s to 1990s

2. The relative skills and economic performance of immigrants have declined.

In 1960, the average immigrant man living in the United States actually earned about 4 percent more than the average native man. By 1998, the average immigrant earned about 23 percent less. (See Table 1.)

The worsening economic performance of immigrants is partly due to a decline in their relative skills across successive waves. The newest immigrants arriving in the country in 1960 were better educated than natives at the time of arrival; by 1998 the newest arrivals had almost two fewer years of schooling. As a result of this growing disadvantage in human capital, the relative wage of successive immigrant waves also fell. At the time of entry, the newest immigrants in 1960 earned 13 percent less than natives; by 1998, the newest immigrants earned 34 percent less.

In short, there has been a precipitous decline — relative to the trend in the native population — in the average skills of the immigrant flow reaching the United States. This historic change in the skill composition of the immigrant population helped rekindle the debate over immigration policy, and is the source of many of the symptoms of immigration that are stressed in this debate.

3. Immigrant earnings will continue to lag behind.

The worsening economic performance of immigrants at the time of entry might not be a cause for concern if their economic disadvantage diminished over time, as immigrants assimilated and acquired skills valuable to American employers. However, the economic gap between immigrants and natives may not narrow substantially during the immigrants’ working lives.

Historical experience suggests that the process of economic assimilation — the acquisition of skills, such as English language proficiency, that American employers value — narrows the wage gap between immigrants and natives by about 10 percentage points in the first two decades after arrival. Immigrants who entered the United States after 1980 had a 25 percent wage disadvantage at the time of entry. If these recent arrivals experience the same rate of economic assimilation as earlier waves, the wage gap between recent immigrants and natives will remain at about 15 percentage points throughout much of the immigrants’ working lives. Put differently, the most recent immigrant waves will probably suffer a substantial economic disadvantage for decades to come.

Table: The Changing Skills of the Immigrant and Native Populations, 1960-1998

4. National origin matters.

Prior to 1965, immigration was guided by the national-origins quota system, which granted visas mainly to persons originating in Western European countries, particularly Great Britain and Germany. The 1965 Amendments to the Immigration and Nationality Act repealed the national origin restrictions, increased the number of available visas, and made family ties to persons already living in the United States the key factor that determines whether a visa applicant is admitted into the country.

As a consequence of these shifts and of major changes in economic and political conditions in the source countries, there was a substantial change in the national origin mix of the immigrant flow. Over two-thirds of the legal immigrants admitted during the 1950s originated in Europe or Canada, one-quarter in Latin America, and only 6 percent in Asia. By the 1990s, only 17 percent of the immigrants originated in Europe or Canada, almost half in Latin America, and 30 percent in Asia. (See Figure 2.)

There are huge differences in economic performance among national origin groups. Immigrants from El Salvador or Mexico earn 40 percent less than natives, while immigrants from Australia or South Africa earn 30 to 40 percent more. These differences in economic performance partly mirror the dispersion in skills across the populations of the source countries. Immigrants who originate in countries that have abundant human capital and higher levels of per-capita income tend to do better in the United States.

The strong link between national origin and economic performance raises an important — and disturbing — problem for immigration policy. Because national origin and immigrant skills are so closely related, any attempt to change one will inevitably change the other.

5. Immigration harms the economic opportunities of the least skilled natives.

Immigrants cluster geographically in a small number of cities and states. This geographic clustering suggests that one may be able to measure the impact of immigration on the labor market opportunities of native workers by comparing natives who reside in immigrant cities (such as San Diego) with natives who reside in cities where few immigrants live (like Pittsburgh). The available evidence indicates that these "spatial correlations" are extremely weak: if one city has 10 percent more immigrants than another, the native wage in the city with more immigrants is perhaps 0.2 percent lower. This finding has led many observers to conclude that immigration has little impact on native employment opportunities.

However, a weak spatial correlation does not necessarily indicate that immigrants have a numerically inconsequential impact on the well being of native workers. Suppose, for example, that immigration into California lowers the earnings of natives in California substantially. Native workers are not likely to stand idly by and watch their economic opportunities evaporate. Many will move out of California into other regions, and persons who were considering moving to California will now move somewhere else instead. These native population flows effectively diffuse the adverse impact of immigration on California’s labor market over the entire economy. In the end, all native workers are worse off from immigration, not simply those who happened to live in the areas where immigrants clustered.

There is evidence that the flows of native workers within the United States — as well as the flows of native firms looking for cheap labor — have indeed responded to immigration. Because of these responses, the labor market impact of immigration must be measured at the national level, rather than at the local level.

Between 1980 and 1995, immigration increased the number of high school dropouts by 21 percent and the number of persons with at least a high school diploma by only 4 percent. During that time, the wage of high school dropouts relative to that of workers with more schooling fell by 11 percentage points. The disproportionate increase in the number of workers at the bottom end of the skill distribution probably caused a substantial decline in the relative wage of high school dropouts, accounting for perhaps half of the observed drop.

Graph: The Changing National Origin Mix of Immigrants


6. Immigration has a severe fiscal impact on the affected states.

In 1970, immigrants were slightly less likely to receive public assistance than natives. By 1998, immigrants had a much higher chance of receiving welfare: Almost a quarter of immigrant households were receiving some type of assistance, as compared to 15 percent of native households.

Two distinct factors account for the disproportionate increase in welfare use among immigrant households. Because more recent immigrant waves are relatively less skilled than earlier waves, it is not surprising that more recent immigrant waves are also more likely to use welfare than earlier waves. In addition, the welfare use of a specific immigrant wave increases over time (both in absolute numbers and relative to natives). It seems that the assimilation process involves not only learning about labor market opportunities, but also learning about the income opportunities provided by the welfare state.

There is little doubt, therefore, that immigrants are making increasing use of public assistance programs. This trend, as well as the expense of providing immigrants with a host of public services has added a new and potentially explosive question to the immigration debate: Do immigrants "pay their way" in the welfare state? A comprehensive study by the National Academy of Sciences concluded that immigration raised the annual taxes of the typical native household in California by about $1,200 a year. The fiscal impact of immigration on the affected states can be quite severe.

7. The net economic gains from immigration are small.

Immigrants increase the number of workers in the economy. Because of the additional competition in the labor market, the wage of native workers falls. At the same time, however, native-owned firms gain because they can now hire workers at lower wages; and many native consumers gain because the lower labor costs lead to cheaper goods and services. As with foreign trade, the gains accruing to the persons who use or consume immigrant services exceed the losses suffered by native workers, and hence society as a whole is better off. However, all of the available estimates suggest that the annual net gain is astoundingly small, less than .1 percent of GDP. In the late 1990s, this amounted to a net gain of less than $10 billion a year for the entire native population, or less than $30 per person.

Immigration, however, does more than just increase the total income accruing to natives. Immigration also induces a substantial redistribution of wealth, away from workers who compete with immigrants and toward employers and other users of immigrant services. Workers lose because immigrants drag wages down. Employers gain because immigrants drag wages down. These wealth transfers may be in the tens of billions of dollars per year.

These facts suggest a new prism for interpreting the immigration debate. Immigration is an income redistribution program, a large wealth transfer from those who compete with immigrant workers to those who use immigrant services or buy the goods produced by immigrant workers. The debate over immigration policy, therefore, is not a debate over whether immigration increases the size of the economic pie in the United States. Rather, it is a debate over how the pie is split.

8. Ethnic skill differentials may persist for at least three generations.

In 1998, 11 percent of the U.S. population was "second-generation" — born in the United States but with at least one foreign-born parent. By the year 2050, the share of second-generation persons will increase to 14 percent, and an additional 9 percent will be composed of the grandchildren of current immigrants. The economic impact of immigration obviously depends not only on how immigrants adapt, but also on the adjustment process experienced by their offspring.

The historical experience of the children and grandchildren of the First Great Migration provides important lessons about the long-run consequences of immigration. A 20 percent wage differential between two immigrant groups in 1910 implied a 12 percent wage differential in the second generation, and a 5 percent wage differential in the third. In rough terms, about half of the average skill differential between any two groups in the first generation persists into the second, and half of the differential remaining in the second generation persists into the third.

The historical lesson is clear: Skill differentials found among today’s immigrants become the skill differentials found among tomorrow’s ethnic groups. If past history is any guide, national origin will still determine the economic performance of the grandchildren of the Second Great Migration at the end of the 21st century. In short, ethnicity matters in economic life, and it matters for a very long time.

9. Ethnic spillovers influence the social mobility of immigrant groups.

The characteristics of the ethnic environment, which I call "ethnic capital," influence the skills and economic performance of the children in the ethnic group, above and beyond the influence of the parents. These characteristics include the culture, attitudes, and economic opportunities that permeate the ethnic networks.

Exposure to an advantaged ethnic environment — in the sense that the environment has abundant human capital — has a positive influence on the children in the group, while exposure to a disadvantaged environment has a negative influence. Ethnic capital creates a type of stickiness in the process of social mobility, making it difficult for persons in disadvantaged ethnic groups to move up and for persons in the advantaged ethnic groups to move down.

The available evidence suggests that the impact of ethnic capital on social mobility may account for half of the persistence in ethnic skill differentials across generations. Because the children of particular ethnic groups tend to follow in the group’s footsteps, ethnic capital effectively lowers the flame under the melting pot from a "full boil" to a "slow simmer."

10. Ethnic ghettoes incubate ethnic differences and slow down the melting pot.

The segregation of young African-Americans into poor neighborhoods helped create the black underclass by depriving these young men and women of particular types of role models, and of the social and economic contacts that provide valuable information about many types of economic opportunities.

There is also a great deal of residential segregation among ethnic groups in the United States. Persons of Mexican ancestry live near other persons of Mexican ancestry, persons of Italian ancestry live near other persons of Italian ancestry, and so on. Because few of the persons in disadvantaged ethnic groups can afford to escape the ethnic ghetto, these enclaves make it easy for ethnic capital to influence the social mobility of the persons who reside in those ghettoes, and help to perpetuate the socioeconomic differences observed across ethnic groups from generation to generation.

Implications for Immigration Policy

So what should the United States do? It is important to get one piece of common sense out of the way quickly: the "top ten" symptoms of immigration listed above — by themselves — have no policy implications whatsoever.

This approach to the immigration debate goes against the grain of most discussions of social policy reform. Typically, those who argue over the parameters of a particular social policy take sides by grasping onto a specific fact, and from that fact they immediately infer some policy reform that the country should pursue. In my view, this way of thinking about the evolution of social policy is just plain wrong.

To see why, let’s single out a particular symptom of immigration over which there is little disagreement: Immigrant use of welfare is high. The policy implication of this fact depends crucially on what the United States is trying to accomplish. If the goal of immigration policy were to ensure that immigration did not place a fiscal burden on the native population, this symptom would surely imply that the United States should take steps to restrict the entry of potential welfare recipients. If the goal of immigration policy were to help the poorest persons in the world, this symptom has no policy relevance — it is the price that the country is willing to pay to achieve a particular humanitarian objective. And the country would want to accept even more immigrants who could become potential welfare recipients.

In the end, a debate over the policy implications of what is known about the economic impact of immigration cannot be based on the evidence alone. Any policy discussion requires explicitly stated assumptions about what constitutes the national interest.

Of course, defining the national interest when it comes to immigration policy is very difficult (and very contentious), even when the debate is restricted purely to the economic issues that tend to frame the immigration debate. To see why, divide the world into three distinct constituencies: the current population of the United States ("natives"), the immigrants themselves, and those who remain in the source countries. To draw policy conclusions from the symptoms of immigration, one has to know whose economic welfare the United States should try to improve when setting policy — that of natives, immigrants, the rest of the world, or some mix thereof. The policy implications implied by the symptoms depend crucially on whose interests the United States cares most about.

By framing the issue in this fashion, the tradeoffs implicit in immigration policy are made crystal clear. The native population probably benefits a great deal when the United States admits high-quality scientific manpower, but the people left behind in the source countries probably lose a lot. Similarly, immigrants benefit when immigration policy favors the entry of their relatives, but natives may lose because this policy does not screen the new entrants and might let in many persons who qualify for social services.

Suppose that the goal of immigration policy were to maximize the economic well being of the native population. And suppose that native economic well being depends both on per-capita income and on the distribution of income in the native population. In effect, the United States wants immigration to make the country wealthier, but it does not want immigration to greatly increase the amount of inequality in the society.

How many and which types of immigrants should the country then admit? The evidence presented in Heaven’s Door, I conclude, can be used to make a strong case that the United States would be better off by adopting an immigration policy that favored skilled workers. And a plausible argument can also be made that the country would be better off with a slight reduction in the number of immigrants.

Immigration policy is essentially a formula that gives "points" to visa applicants on the basis of various characteristics, and then sets a passing grade. The variables in the formula determine which types of persons will be admitted, while the setting of the passing grade determines how many persons will be let in. In 1965, the United States began to use an admissions formula that, to a large extent, had only one variable, indicating if the visa applicant had a family member already residing in the United States. The immigrant flow would be much more skilled if the United States, like other immigrant-receiving countries, expanded the admissions formula to include not only family connections, but such other factors as education, age, and English proficiency.

And how many immigrants should the country admit? In theory, the United States should admit an immigrant as long as the contribution made by the immigrant exceeds the costs imposed by the immigrant. It is difficult to come up with the "magic number" unless one becomes much more specific about how to balance the facts that immigration increases the size of the economic pie and changes how the pie is split. Those who care about economic efficiency will typically want to see more immigrants: The greater the number of immigrants, the greater the gains to employers (through lower wages) and consumers (through lower prices). Those who care about distributional issues will typically want to see fewer immigrants: The greater the number of immigrants, the greater the dislocation in labor markets, and the greater the losses suffered by those who compete with immigrant workers.

I suspect that an annual flow of 1 million immigrants is probably too large — regardless of whether the losers are at the bottom or at the top of the skill distribution. Such a large flow can substantially depress the economic opportunities of workers who compete with immigrant labor. A good place to start the process of converging to the "magic number" might be to let in 500,000 immigrants per year — which happens to roughly correspond with the recommendation made by the Commission for Immigration Reform in 1997.

In the end, the objective of immigration policy will reflect a political consensus that inevitably incorporates the conflicting social and economic interests of various demographic, socioeconomic, and ethnic groups, as well as political and humanitarian concerns. Perhaps after debating these issues, the American people will place the economic concerns aside, and choose an immigration policy that stresses the humanitarian or political consequences of immigration.

Such a resolution of the debate does not diminish the value of entering a discussion that informs the country about the economic consequences of immigration. A wise decision requires that the American people be fully aware of the price they will have to pay if the United States chooses to adopt an immigration policy that minimizes or ignores economic considerations.

Dr. Borjas is the Pforzheimer Professor of Public Policy at the John F. Kennedy School of Government, Harvard University.