April 25, 2006
"Immigration: The Economic Impact"
held by the
United States Senate, Committee on the Judiciary
April 25, 2006
Barry R. Chiswick*
UIC Distinguished Professor
Department of Economics
University of Illinois at Chicago
Institute of Government and Public Affairs
University of Illinois
Program Director, Migration Studies
IZA - Institute for the Study of Labor
web address: www.uic.edu/~brchis
* The ideas presented in these comments are those of Prof. Chiswick and not to be attributed to any institution with which he is affiliated. Data on immigration flows are from the U.S. Department of Homeland Security, Yearbook of Immigration Statistics: 2004, Washington, D.C.: U.S. Department of Homeland Security, Office of Immigration Statistics, 2006.
Immigration: The Economic Impact
Barry R. Chiswick
1. Impacts on Relative Wages
2. The "Need" for Low-Skilled Immigrants
3. A Century Ago
4. Fallacies in Estimating Immigrant Impacts
5. Does Country of Origin Matter?
6. Immigration Law and Low-Skilled Immigrants
7. Border and Interior Enforcement
8. The Current Immigration System
9. How Many Immigrants?
Immigration: The Economic Impact
Barry R. Chiswick
The theme of this hearing is "Immigration: The Economic Impact." When I am asked the question "What is THE economic impact of immigration?" where the tone indicates the emphasis on the word "the," I respond that this is not be best way to couch the question. There are two fundamental questions. One is: "What is the optimal size of the immigrant population?" The other is: "What are the different impacts of immigrants that differ in their productivity-related characteristics?"
Impacts on Relative Wages
Let us begin with a discussion of the second question. Conceptually, it is best to think in terms of two types of immigrants, which for simplicity we will call high-skilled and low-skilled, with the same two skill groups represented in the native-born population. High-skilled immigrants will have some characteristics in common, without regard for their country of origin. They tend to have high levels of schooling, which means they tend to have a high degree of literacy, perhaps also numeracy, critical thinking or decision-making skills. Many, but not all, will have a high degree of scientific or technical knowledge, and in the modern era a high comfort level with computer technology. Many, but certainly not all, will either have a degree of proficiency in the destination language (in this case, English) or the ability to acquire proficiency in that language shortly afer arrival. These are all characteristics that have been shown to improve the earnings of immigrants and to facilitate their economic adjustment in the host country.
Although particular individuals may differ, low-skilled immigrants generally have little formal schooling, limited literacy proficiency in their mother tongue (the language of their origin country), and limited scientific and technical knowledge. These are characteristics associated with low earnings in the destination.
High-skilled and low-skilled immigrants will, in general, have different impacts on the host economy and labor market. Labor markets behave in a manner similar to other markets, in that a greater supply of a given type of labor tends to depress the market wage of workers with similar characteristics. An increase in the supply of a given type of worker also increases the productivity of the complementary factors of production with which it works, including other types of labor and capital. To give a simple example, an increase in the supply of low-skilled restaurant kitchen help will result in more competition for this type of job and lower wages for ordinary kitchen workers. Yet this will increase the productivity (and hence wages) of the master chefs because with more help for the menial kitchen chores they can spend their time on the highly specialized tasks for which they have trained. By the same token, an increase in the supply of high-skilled chefs would raise the productivity of low-skilled restaurant kitchen workers since they would have more master chefs for whom to work.
The result of high-skilled immigration tends to be an increase in the wages of all low-skilled workers (and reduce their use of public income transfers) and a decrease in the wages of high-skilled natives. This reduces income inequality, which we generally view as a good development. Like high-skilled natives, the taxes paid by high-skilled immigrants tend to be greater than the costs they impose on the public treasury through the income transfers they receive, the schooling received by their children, and the publicly subsidized medical care that they receive. High-skilled immigrants are also more likely to bring with them the scientific, technical and innovative skills that expand the production capabilities of the economy. As a result, the population as a whole tends to benefit from high-skilled immigration, although with some benefiting more than others.
Now consider the impacts of low-skilled immigration. While these immigrants tend to raise the earnings of high-skilled workers, their presence in the labor market increases competition for low-skilled jobs, reducing the earnings of low-skilled native-born workers. This not only increases income inequality, which is rightly considered to be undesirable, it also increases the need among low-skilled natives for public assistance and transfer benefits. Because of their low earnings, low-skilled immigrants also tend to pay less in taxes than they receive in public benefits, such as income transfers (e.g., the earned income tax credit, food stamps), public schooling for their children, and publicly provided medical services. Thus while the presence of low-skilled immigrant workers may raise the profits of their employers, they tend to have a negative effect on the well-being of the low-skilled native-born population, and on the native economy as a whole.
These points are not purely theoretical arguments. In the past two decades the real wages of low-skilled workers have remained stagnant even as the real earnings of high-skilled workers have risen. As a result, income inequality has increased. Several factors have been responsible for this development, but one of them has been the very large increase in low-skilled immigration.
The "Need" for Low-Skilled Immigrants
"But," I am often asked, "don't we need low-skilled immigrant workers to do the jobs that native workers are unwilling to do?" I respond: "At what wage will native workers decline to take these jobs?" Consider the following thought experiment: What would happen to lettuce picking or the mowing of suburban lawns if there were fewer low-skilled workers? Earlier this month on ABC's Nightline program a winter lettuce grower in Arizona provided the answer. He acknowledged that he would pay higher wages to attract native-born workers and he would speed up the mechanization of lettuce harvesting. The technology is there, but with low wages for lettuce pickers there is no economic incentive for the growers to mechanize or invest in other types of new technology. If the supply of low-skilled immigrant workers decreased substantially, mechanical harvesting would replace many of them with capital (machines) and more highly paid native workers. How would suburban lawns get mowed if there were fewer low-skilled immigrant workers? Wages for lawn care workers would surely rise. The result would be that more teenagers and other low-skilled native workers would find it worth their while to make themselves available for this work.
In addition to this substitution of one type of labor (youthful and low-skilled natives) for another (low-skilled immigrants), there would be other adjustments to the higher cost of lawn mowing. One would be letting the grass grow longer between mows - say, every ten days instead of weekly. Another would be the substitution of grass that grows more slowly, or the substitution of ground cover or paving stones for grass, etc. The point is that there would be many ways for consumers and employers/producers to respond to the higher wages of low-skilled workers to mitigate the adverse effects of having fewer low-skilled immigrants.
A Century Ago
At this point in the conversation, someone usually points to the period of mass immigration of unskilled workers from the 1880s to the 1920s: If these arguments are valid now, wouldn't they have applied at that time as well? - and we know that immigration was a tremendous net benefit to the United States at that time. The answer is both yes and no. The economy and economic institutions of 100 years ago were quite different from those of today in ways that are both important and relevant to our discussion. Then, rapid industrialization of the American economy generated a very large demand for unskilled workers in mines and in factories producing everything from steel to shirts. This is no longer the case. Technological change, the increased cost of even low-skilled labor (wages plus fringe benefits and employment taxes), the falling cost of capital equipment, and globalization/international trade have sharply reduced the demand for low-skilled workers in U.S. manufacturing, mining, agriculture, and even service occupations and industries. Moreover, 100 years ago income inequality and income distribution issues were not a matter of public policy concern. If there were poor people in the United States - so be it. If private individuals and charities helped the poor - fine, but there was nothing like the tax-funded income transfer system in place today.
Yet in some ways the mass immigration from Europe 100 years ago had a similar impact as the one we are facing today. By holding down the wages of low-skilled workers in the industrializing centers of the economy, especially in the Northern states, rural-urban and South-North migration was slowed. Rural and Southern poverty persisted longer than they might have otherwise, and it was only after war (WWI) and immigration restrictions (in the 1920s) had effectively stopped the European migration that these poverty-reducing internal migrations resumed. While there is no question that there were long-term benefits from the massive wave of immigrants for the country as a whole, it is also true that the low-skilled native-born workers of that time paid a price.
Fallacies in Estimating Immigrant Impacts
In the course of these hearings on the economic impact of immigration, you may receive testimony regarding a body of literature that attempts to estimate this impact. In this literature a statistical technique, regression analysis, is used to show how the wages of native workers (or low-skilled natives in particular) in a state or metropolitan area are affected by the extent to which there are immigrants (or low-skilled immigrants) in the same area. These studies tend to find no relation, or sometimes a very small relation, between the presence of immigrants and wage levels.
There is nothing wrong with regression analysis per se as a statistical technique, but its application in this case is flawed. This application of regression analysis requires us to assume that each state or metropolitan area is a self-contained economy, with little or no in-and-out movement of workers, of capital, or even of goods and services. We know, however, that this is not the case. Labor, capital and goods are highly mobile across state boundaries and metropolitan areas. What we learn from these studies is not that immigrants have no effect on wages, but that these wage effects - whatever they may be - have spread throughout the country. Although it does provide evidence that markets in the United States function quite efficiently, the impacts of immigration can not be detected by this statistical technique.
At the aggregate level, many analyses consider immigrants as an undifferentiated whole without distinguishing between high-skilled and low-skilled workers. These also provide misleading implications, often to the effect that immigrant impacts on wages and income distribution are small. When the positive economic benefits of high-skilled immigration are lumped together with the more negative consequences of low-skilled immigration, they appear to cancel each other out because there are both gains and losses. In the real world, however, the penalty paid by low-skilled natives because of high levels of low-skilled immigration is not so easily cancelled out by the positive impacts of high-skilled immigration.
Does Country of Origin Matter?
To this point I have not said anything about country of origin. That is because country of origin per se is not really relevant for an analysis of economic impacts. What is most relevant is the skills that immigrants bring with them.
Immigration Law and Low-Skilled Immigrants
I have also not said anything yet about legal status. For various reasons, most individuals working in the United States in violation of immigration law are low-skilled workers. But most low-skilled workers are not "undocumented" aliens. Most low-skilled workers were born in the United States and hence are citizens by birth.
Current U.S. immigration law, however, encourages the legal immigration of low-skilled workers. This encouragement comes through the kinship preferences for various relatives built into our legal immigration system and to the smaller diversity visa program. Our immigration law permits a "snowball effect" where even immigrants granted a visa for the skills they bring to the U.S. labor market can sponsor low-skilled relatives who will then legally work in the U.S.
Of the 946,014 people who received Permanent Resident Alien visas in 2004, 65.6 percent entered under one of the several kinship categories, 8.8 percent entered as refugees or asylees, 5.3 percent entered under "diversity" visas, and 3.5 percent had a cancellation of deportation order. The 155,330 employment-based visas represented only 16.4 percent of the total. However, only about half of those who received an employment-based visa were themselves skill-tested (less than 73,000), while the remainder of these visas were received by their spouses and children. Thus, only about 7.6 percent of the nearly one million visa recipients were asked a question about their skills. (See Figure 1.)
The 1986 Immigration Reform and Control Act (IRCA) was sold to the American public as having two major features - amnesty which was to "wipe the slate clean" of undocumented workers, and employer sanctions which was to "keep the slate clean" - along with some increased border enforcement of the immigration law. Employer sanctions were intended to cut off the "jobs magnet" that attracted undocumented workers to the United States. Half of the political bargain was fulfilled. Under its two major amnesty provisions legal status was granted to nearly 3 million undocumented individuals, nearly all of whom were low-skilled workers, and millions more have subsequently been able to immigrate as their relatives. It is noteworthy that while in 1986 the word "amnesty" was used outright, in the current political debate the "A" word is anathema to the proponents of what is euphemistically called "earned legalization." This by itself is testimony to public perception of the failures of the 1986 Act.
Border and Interior Enforcement
Border enforcement, both at land borders and at airports, is a necessary element in the enforcement of immigration law. Border enforcement by itself has not, can not, and will not work in controlling illegal entry of undocumented immigrants. If a potential immigrant is unsuccessful in penetrating the border on the first try, success may be had on the second or third try. This may be done by "entry without inspection" (i.e., sneaking across the border) or by using "fraudulent documents" at a border crossing point. Alternatively, a "visa abuser" enters into illegal status by violating a condition of a legally obtained visa - by working while on a tourist visa, for example, or by overstaying the time limit permitted on a temporary visa.
Thus, border enforcement must be complemented with "interior enforcement." The 1986 Act focused on "employer sanctions," penalties for employers who knowingly hire people who do not have the legal right to work in this country. There has, however, been no serious effort over the past two decades to enforce employer sanctions. Modern technology makes it easier to create fraudulent documents, but it also makes it easier to develop more stringent identity checks. There are two major failings in the current system. Employers are not given a "foolproof" mechanism to readily identify those with a legal right to work, and the Federal authorities show no interest in enforcing the law, except for an occasional "show raid."
It is not obvious that new enforcement legislation (e.g., to criminalize an illegal status) is called for. What is obvious is that illegal immigration can not be controlled without a political will to enforce current immigration law. This includes providing employers with a simple and "foolproof" mechanism for identifying workers with a legal right to work in the U.S. along with more stringent enforcement of employer sanctions.
The Current Immigration System
The current legal immigration system is not serving the best economic interests of the United States. Only a small percentage of the immigrants who enter the U.S. legally in any year (less than 8 percent) are screened for their likely economic contribution to this economy. The vast majority enter under a nepotism system (the kinship preferences), with a smaller group entering under a lottery (diversity visas). To enhance the competitiveness of the U.S. economy in this increasingly globalized world, where efficient competitors are emerging across the world, the U.S. needs to change the basic question from "To whom are you related?" to "What can you contribute to the U.S. economy?"
Other highly-developed democratic countries - Canada, Australia, New Zealand - introduced "skills-based" immigration policies several decades ago. More recently, some countries in Western Europe have done the same. Some, like Canada and Australia, use a "points system" in which points are awarded based on characteristics that research has shown to enhance the earnings of immigrants such as age, schooling, technical training, and proficiency in the host country's language. Those with more than the threshold number of points receive a visa for themselves, their spouse and their minor accompanying children. This shift in emphasis in the rationing of visas would increase the skill level of immigrants and provide greater economic benefits to the U.S. economy than the current system.
A points system has many advantages over the current targeted employment-based visas. Under the current system a complex and very expensive bureaucratic process is required for employers to demonstrate to the U.S. Department of Labor not only that the visa applicant is qualified for a specific job but also that there is no qualified person with a legal right to work in the U.S. who will take the job at "prevailing wages." Even then, the worker who obtains a visa through this process is not obliged to remain on that job or with that employer.
Other proposals would use market mechanisms to "close the gap" between the large demand for visas and the much smaller supply that the U.S. is willing to make available. One possibility would involve auctioning visas; another involves charging a large market-clearing "visa fee." Among other advantages of these market mechanisms is that people in the U.S. can express their preferences for bringing relatives and friends by contributing to the price of their visa. Nor does there need to be only one mechanism - a skill-based system and a market-based system could both be used.
How Many Immigrants?
This returns us to a question posed early in this testimony: "What is the optimal size of the immigration flow?" The optimal immigration policy is neither a completely open door nor a completely closed one. There is no magic number or proportion of the population. Currently, legal immigration is running at approximately one million immigrants per year. This is on a par with the peak period of immigration from 1905 to 1914, when immigration also averaged one million per year. Yet, relative to the size of the U.S. population, current legal immigration is about one-fourth of the ratio in this earlier period. There is no clear evidence that the U.S. has exceeded - or even reached - its absorptive capacity for immigration. The U.S. economy and society exhibits a remarkable adaptability to immigrants, and thus far immigrants continue to show considerable adaptability to the U.S. economy and society. This adaptability means that the U.S. economy can absorb a continuous stream of immigrants without fracturing the system.
The demand for visas to enter the U.S. is very strong and, if anything, it seems to be increasing. This is a credit to the U.S. economy, society, and political system. The number of visas the U.S. political process is willing to supply is not immutable. The greater the economic benefits of immigration, the larger the optimal number of visas and the greater the willingness of the American public to provide them.
A comprehensive immigration policy reform would reduce undocumented migration by more stringent enforcement of existing law. It should also include the adoption of a skill-based points system and/or market mechanisms to ration visas, while limiting kinship migration to the immediate relatives of U.S. citizens (spouse, minor children, aged parents). These policies would increase the benefits of immigration for the American public, providing economic incentives to increase the supply of visas and hence the annual total number of immigrants entering the country legally.