The immigration bill about to be signed into law by President Bush constitutes only the fourth major shift in U.S. immigration policy in a century. The first of those changes was the enactment by Congress in the late 19th century of the first restrictions upon settlement in the U.S. by particular classes of persons, restrictions that eventually led to the banning of immigration from most Asian countries. The 1920s witnessed a second shift -- the introduction of the national-origins quota system, a visa allocation formula that greatly limited the number of immigrants from countries in Southern and Eastern Europe.

The third shift, in 1965, rid immigration policy of national-origin restrictions and made family ties with U.S. residents the main determinant for entry. The fourth, just enacted, greatly increases the size of the immigration flow into the U.S. -- to 700,000 from 500,000 a year, not counting refugees. Skilled workers (and their families) will be allocated 140,000 visas a year, up from 86,000 under current law. In other words, almost half of the additional visas made available by the new law will go to skilled workers.

The increasing emphasis on skills as a way of allocating visas among the many applicants will have a beneficial impact on the U.S. economy in the next decade. The gains will include: an increase in tax revenue, a reduction in the costs of social services and an increase in the supply of skilled workers. The new legislation makes the U.S. more competitive in the international marketplace for human capital. But because the basic structure of U.S. immigration law remains what it has been since 1965, it is unlikely that immigration in the 1990s will be as profitable for the U.S. economy as immigration was in the 1950s and 1960s. To see why, look at the economic impact of immigration in recent years.

More and more of America's unskilled workers are immigrants. Immigrants accounted for just 12% of all high school dropouts (persons with less than a high school education) in the U.S. labor force in 1975. By 1985, the proportion of high school dropouts who were immigrants had almost tripled -- to 32%.

In contrast, the contribution of immigrants to the supply of skilled workers remained constant over the period. In both 1975 and 1985, 6% of all college graduates in the workforce were immigrants.

Because immigrants are increasingly likely to be unskilled, they have not been doing as well as they used to in the labor market. In 1960, the average hourly wage of immigrants who had been in the country fewer than five years was 12% lower than that of natives. By 1970, the wage gap between immigrants who had been in the country for fewer than five years and natives had climbed to 15%. In 1980 the wage gap between the most-recently arrived immigrants and natives was 26%.

This spreading gap has been felt through the whole economy. If the immigrants who arrived in 1975-1979 had been as skilled as those who arrived in the early 1960s, U.S. GNP would be about $6 billion higher this year and every other year of those immigrants' working lives. These losses will mount as more unskilled immigrant waves enter the country. If the skills of the immigrants who arrived in the 1980s did not improve over those of the immigrants who arrived in the 1970s (a question that cannot be answered until census data becomes available in 1993), potential GNP for each and every year of those immigrants' working lives will be reduced by an additional $12 billion yearly. The tax revenues forgone because of the skills drop-off is also substantial: about $4 billion a year.

The decline in immigrants' skills aggravates the burden of welfare and social programs. Although the conventional wisdom is that immigrants shy away from welfare, the facts are quite different. Both immigrants and natives became more prone to take welfare in the 1970s, but the rate of increase was much faster for immigrants than for natives. In 1970, immigrants were slightly less likely than natives to take welfare: 5.9% of immigrant households and 6.1% of native households were receiving welfare. By 1980, immigrants were more likely than natives to take welfare: 8.8% of immigrant households as against 7.9% of native.

Over the life cycle of those immigrants, the present value of the welfare costs associated with the typical household that immigrated in the 1970s will be $12,746 (in 1989 dollars), while the lifetime welfare costs of the typical household that immigrated in the 1950s will be $7,178. The welfare costs of the typical native household of the 1970s will be $7,909. If the 1.6 million immigrant households who entered the country in the 1970s had been as averse to welfare as the immigrant households who arrived in the 1950s, the nation's welfare bill would have been cut nearly in half: to $11.5 billion over those immigrants' working lives, from $21 billion.

The deteriorating economic performance and increasing attachment to the welfare system of immigrants is not solely the result of an influx of less-educated workers. More recent immigrant waves simply do not perform as well as natives with similar levels of education. The average hourly wage gap between newly-arrived immigrants and natives with the same education and age was only 9% in 1960. By 1970, newly arrived immigrants were earning 14% less than demographically comparable natives; by 1980, the gap had increased to 20%.

One possible explanation for this trend is that recent immigrant waves have been originating in different countries than the earlier waves did, and there is substantial dispersion in the economic performance of national origin groups in the U.S. Newly-arrived immigrants in 1980 (members of the 1975-1979 wave) from Britain or France earned 15% to 20% more than demographically comparable natives, while those from Korea, Mexico or the Dominican Republic earned 20% to 30% less.

Not all skills are alike. In general, persons originating in countries that most resemble the U.S. have skills which are most easily transferable to the U.S. labor market. It is not surprising, therefore, to find that immigrants from highly-developed, industrialized countries outperform immigrants who originate in developing economies. Because immigration policy does not discriminate on the basis of national origin (and rightly so), the awarding of visas to "skilled" workers ignores the extent to which skills are transferable to the U.S. economy, and dilutes the effect of the legislation.

It is easy to misinterpret the implications of these facts. Immigration provides many benefits for the U.S. But because of the immigration policies in place since 1965, it is less economically lucrative than it used to be.

The 1990 immigration legislation can be interpreted as an attempt to increase the economic benefits from immigration. During the 1980s, the economic return on skills grew in the U.S. The wage gap between the earnings of college graduates and less-educated workers swelled to postwar records. Higher wages for skilled work in the U.S. created new incentives for the skilled workers of other countries to migrate to the U.S.

Until recently, however, U.S. immigration policy was a major obstacle for the entry of these persons -- skilled workers had trouble entering the country unless they had relatives already living here. Other countries participating in the immigration market, principally Canada and Australia, bid these workers away.

It is undoubtedly true that the U.S. has profited greatly from immigration in the past. In any case, it is far from clear that the goal of immigration policy should be to maximize economic gain to natives. Whatever the ultimate objectives, however, it is also clear that a policy which ignores the economic consequences of immigration can impose substantial costs on the U.S.

---