There has been a long-standing interest in the flows of persons across countries and in particular the international flow of skilled migrants. There are two major components in the literature on international migration, each of which is concerned with the skill composition of international migrants. One strand of literature focuses on the impact of immigration on the receiving-country economy. Because the impact of immigration depends on the skill composition of immigrants, understanding the determinants of the skills of new immigrants is an important element of this research program. Most of the analyses of the determinants of the magnitudes and skill composition of U.S. immigrants use data on the foreign born from the U.S. Census Bureau.1 However, this literature has two major shortcomings. First, Census data do not provide information on entry visas. The U.S. immigration system selects immigrants according to a wide variety of criteria, including mainly family relationships to U.S. citizens. And some foreign born are temporary migrants or migrants without a legal basis for staying or working in the United States. This heterogeneity in selection criteria and the constraints on immigration associated with ceilings on country and visa categories are not typically taken into account in the analyses, and thus little is known about immigrants who are selected by skill. Recently available survey-based data on legal immigrants, moreover, indicate that there are substantial educational differences across immigrants chosen by different criteria.2
The second shortcoming of the literature focusing on U.S. immigration is that the analytical framework used is based on a model originally designed to [End Page 57] analyze the choice of occupations by workers of different skill levels in a domestic economy in which there are no differences in the rewards to skill.3 The framework is thus not well suited for studying international migration in a world in which there are large differences in earnings for workers of the same skill. In addition, given the importance of family connections, job networks, and constraints on the number of immigrant visas, standard economic models of self-selection have not been wholly successful in providing empirical evidence on worker migration based on data on immigrant flows or the U.S. stocks of foreign born.4
Click for larger view
|Figure 1 |
Annual Number of Foreign Student Visas Issued, by Receiving Country, 2003 and 2004
A second strand of literature is concerned about the impact of skilled migration from the perspective of sending countries. Many of the earlier contributions to this literature, focusing on the outflow of workers and using the rubric "brain drain," were mostly theoretical.5 However, in recent studies global information on the stocks of foreign born by education in developed countries has been [End Page 58] assembled to quantify the magnitude of the outflow of skilled migrants.6 And the recent analytical literature has taken a somewhat more nuanced view of the phenomenon.7 This newer literature, however, has underappreciated how much of the training of individuals born in poor countries occurs in rich countries. Just four receiving countries of international students—the United States, Australia, Canada, and Britain—are currently admitting more than 525,000 students a year. And, as indicated in figure 1, the United States accounts for close to half of this flow.8
The flow of foreign students admitted to the United States is considerably larger than the two other skill-based flows of immigrants and nonimmigrants to the United States. In contrast to the more than 240,000 foreign born who were admitted to the United States in 2004 as formal students coming to augment their skills, only 73,212 permanent immigrants who qualified for an employment visa on the basis of their skills were admitted.9 And only 65,000 foreign workers were admitted as temporary migrants in 2004 on the basis of their skill qualifications within the H1B category.10 Even if half...