Internal migration and regional population change
The U.S. population has typically been a very mobile one, with up to 20 percent of the population changing residence each year, 7 percent changing county of residence from one year to the next, and perhaps 4 percent making interstate moves. More recently (1992), these figures were slightly lower (17, 6, and 3, respectively), and they are likely to decline further in the coming years. Such a decline would reflect the effects of two factors: the age pattern of geographic mobility rates (see Figure III.C.1), and the aging of the population (cf. differences by education).
The presence of return and repeat migration means that mobility is concentrated within certain segments of the population. Thus, for example, during a four-year interval some time back, 58 percent of the population remained in the same residence, whereas if mobility were randomly distributed from year to year the percentage of nonmovers would be 47 with a 17 percent annual migration rate and 41 with a 20 percent migration rate (.47 = .83 to the power 4; .41 = .80 to the power 4).
Regional population change in principle will reflect both differences in rates of natural increase across regions and net population movements (internal migration and immigration) among and to the different regions. As can be seen clearly in Table III.C.1, the first of these two components is relatively minor; internal migration and immigration play the more important role in determining regional population change.
That is, with the exception of the West, there are only modest differences in rates of natural increase by region. Even including the West, the maximum regional difference in natural increase is on the order of 2:1, while the maximum difference in net change is greater than 6:1.
Immigration, which accounts for almost a third of recent U.S. population growth, is heaviest to the West, both absolutely and relatively, and also plays an important role in contributing to population growth in the South and the Northeast. It is internal migration that plays the major role in contributing to regional differences in growth rates: outmigration from the Northeast and inmigration to the South are both very heavy.
Geographic mobility rates have been highest in the West (21 percent in 1992) and the South (19 percent), lower in the Midwest (16 percent), and lowest in the Northeast (12 percent). These regional differences to a large degree reflect differences in the age compositions of the respective regional populations.
An obvious question is whether or not these recent trends reflect new developments or old patterns. Table III.C.2 shows the combined effects of the trends in migration, immigration, and natural increase on changes in the regional population totals from 1980 to 1990. Again, the rapid growth in the South and West is evident: these two regions accounted for almost 90 percent of the total growth in population during the 1980s. For this slightly longer look back, it's also clear that the Northeast and the Midwest have traded places.
Looking further backward reveals additional differences with respect to more recent data. Taking the long view, migration within the U.S. has been predominantly out of the South (especially among blacks) and towards the West. In the past few decades, however, these traditional migration flows have changed somewhat.
In particular, the South emerged as the largest "importer" of people in the 1970s; although challenged by the West in the latter half of the 1980s (largely because of immigration), it has regained its more recent leading position, as may be seen in Table III.C.1. The absence of any substantial domestic migration to the West is a relatively recent phenomenon, reflecting economic difficulties in California (which experienced net domestic outmigration of almost 1.2 million people from 1990-94). For most of the 1960s, 1970s, and 1980s the West experienced high levels of net inmigration.
The Midwest has been a steady net exporter of people over the past three decades, although the volume of outmigration has diminished considerably since the early 1980s (cf., "Rust Belt" and decline of employment in the auto industry). The Northeast has been the most volatile region, with slight net inmigration during the 1960s, net outmigration in the 1970s and early 1980s, a brief spell of net inmigration during the mid/late 1980s, and net outmigration again since the late 1980s.
The high growth taking place in the West and the South is expected to continue, at least in the near term, as may be seen in Figure III.C.2. This contrasts with the slow growth projected for the Northeast and most of the Midwest.
Metropolitan and nonmetropolitan population growth
Migration within the U.S. has traditionally been primarily from nonmetropolitan to metropolitan areas. During the past 40-45 years, overall population growth in the U.S. has been slowing. This has been the pattern within metropolitan areas (which account at present for roughly 80 percent of the nation's population), in which central city populations have been declining (cf., 1992 mobility within metropolitan areas: 38 percent within suburbs, 37 percent within central cities, 17 percent from central cities to suburbs, and 9 percent from suburbs to central cities).
Smaller metropolitan areas (with population under 1.5 million) have experienced the greatest growth. Nonmetropolitan areas, which were the slow-growth places in the 1950s and 1960s (and earlier), experienced relatively rapid growth in the 1970s, but since then their growth has again been surpassed by that of metropolitan areas (note racial differences, as shown in a transparency presented in class; regional differences, as shown by Long and DeAre's Figure 2; and updating via Table III.C.3).
Economic aspects of population redistribution
There are three distinct aspects of population redistribution that we've been reviewing: interregional, metropolitan/nonmetropolitan, and central city/suburban. These aspects are clearly interrelated, and linked to changes in transportation and communication technology as well as to a changing structure of national and international economic organization.
Growth of the Sunbelt is seen as reflecting the emergence first of the West and subsequently the South as major industrial growth poles of the U.S. economy. Lower transportation costs via the interstate highway system facilitated geographic dispersion of production activities. Lower manufacturing wages in a largely nonunion environment (particularly in the South) also contributed to the attractiveness of the Sunbelt.
In addition to these "pull" factors, negative aspects of northern industrial cities -- such as congestion, high union wages, crime, high taxes, and high cost of land -- were argued to "push" jobs and people out of the urban North. Further, the existence of a growing "footloose" retirement population, with income sources such as private pensions and social security that were not tied to being in a specific location, in conjunction with generally lower land, living, and amenity costs were seen as contributing to Sunbelt expansion.
The more rapid growth of nonmetropolitan areas in the 1970s was attributed in part to decentralization of the spatial distribution of jobs in a maturing economy -- i.e., an economy with a growing share of employment in services where ease of transport was no longer relevant. Changes in communication technology further facilitated this process, as did extension of the interstate highway system to more remote rural areas and provision of power lines and public services. Again, the "footloose" retirement population is likely to have contributed to this phenomenon as well (cf., State College).
In effect, it was suggested that cities had exhausted their scale or agglomeration economies. The relative slowdown of nonmetropolitan growth since the 1970s suggests that such economies have not yet been entirely exhausted, but as noted earlier it is the smaller rather than the larger metropolitan areas that have been experiencing more rapid growth.
Finally, the shifts of metropolitan area populations from central cities to suburban rings represent continuation of a phenomenon that has been going on for quite some time. Of note here is the fact that this population redistribution is quite selective by income level, and reflecting correlations of income and race, it has contributed to racial polarization in a number of cases (cf., school segregation and busing).
Immigration: Supply of and demand for immigrants
A starting point for analyzing immigration is to think about the supply of and demand for immigrants. Economic factors are an important aspect influencing the supply of immigrants. Reflecting our general microeconomic model of migration, we would expect earnings differences across countries as well as the costs of migration to be key economic variables (cf., adaptability). In this context, business cycle fluctuations would be expected to have some impact on migration behavior, as perhaps best illustrated by the decline in immigration to the U.S. during the Great Depression of the 1930s (cf., economic activity here and abroad; impact on returns and effect on ability to pay costs of migration).
Further, noneconomic "push" factors are also likely to influence the supply of immigrants. Most notable in this regard are political events such as wars or revolutions that tend to create flows of refugees.
Demand factors for immigrants are reflected in immigration policy, as specified in immigration law. In particular, legislation normally spells out the number of visas available for legal immigrants and the criteria for rationing those visas.
A second aspect of policy reflecting demand for immigrants is the degree of enforcement of legislation. Greater enforcement of immigration restrictions results in greater deterrence to and hence lower levels of illegal immigration.
Consider now changes over time in U.S. immigration policy.
U.S. immigration policy
As Espenshade et al. note, U.S. immigration policy has swung back and forth over time, seeking to achieve different and sometimes conflicting objectives. These objectives have included economic (attracting labor in general, as occurred early in U.S. history, or certain skills), humanitarian (family reunification), cultural (promoting ethnic and racial diversity), and political (welcoming various political refugees or restricting access) considerations.
There were no laws regulating the admission of aliens to the U.S. until the latter part of the 19th century. From that time on, a number of restrictive laws were passed, beginning with the Chinese Exclusion Act of 1882. This general immigration statute prohibited entry of Chinese laborers (who earlier had come to the U.S. to work on construction of the transcontinental railroad) for a period of 10 years, limited naturalization of Chinese, and provided for deportation of illegal Chinese immigrants.
Immigration in the early 20th century was extremely high relative to historical levels. However, significant restrictions on immigration were imposed not long after World War I, beginning first with the Quota Act of 1921. This limited European immigration initially to 3 percent of each country's foreign-born population in the U.S. as of the most recent census. This restriction was subsequently reduced to 2 percent of each nation's number of foreign-born residents as of the 1890 census. This shift had the effect of discriminating against those from Southern and Eastern Europe, who came to the U.S. primarily after 1890.
In 1924 the Immigration and Naturalization Act was passed. The act lowered the annual overall quota from 358 thousand to 154 thousand. The combined effects of the Quota Act and the Immigration Act were to reduce immigration from Southern and Eastern Europe from 45 percent of the total in 1921 to only 15 percent by 1929. The beneficiaries of the act (given the reduction in the overall quota) were immigrants from Great Britain, Ireland, and Germany.
The next significant piece of immigration legislation was the McCarran Act in 1952, which aimed to "keep America pure" and also included "antisubversive" provisions. The annual ceiling on immigration was maintained at roughly 154 thousand, with Europe accounting for almost all of the total (150 thousand). The bulk of the European quotas was allocated to Great Britain (65K), Germany (26K), and Ireland (18K), continuing the favorable treatment these countries had received earlier.
The national quotas first established in the 1920s were legislated out of existence in 1965 with passage of the Immigration and Nationality Act Amendments. The volume of legal immigration was also increased. An overall ceiling of 170K was established for immigrants from the Eastern hemisphere (with a maximum of 20K for any individual country), along with a ceiling of 120K for immigrants from the Western hemisphere (no country ceiling). Subsequent amendments (1978) would abolish the hemispheric ceilings but retain the individual-country ceiling.
The most significant aspect of the 1965 legislation was the fact that, contrary to past policy, it deemphasized skill or occupation requirements for immigrants. Instead, emphasis was placed on family reunification. As we shall see below, this resulted in a marked shift in the composition of the immigrant population in terms of countries of origin and in terms of adaptability to the labor market.
As Borjas notes in his article, the overriding emphasis on family reunification that now characterizes U.S. immigration policy is unique among major immigrant-receiving countries (e.g., Canada, Australia, and New Zealand in addition to the U.S.). Further, a 1980 amendment provided that immediate relatives of U.S. citizens would be accepted outside of the statutory limits on immigration.
In 1986, the Immigration Reform and Control Act (IRCA) was passed. In response to growing concerns about illegal immigration to the U.S., this act explicitly prohibited employers from "knowingly hiring... aliens not authorized to work in the United States." It further provided for sanctions against employers who hire illegal aliens, in the form of fines ranging from $250 to $10,000 per illegal alien.
A second major thrust of IRCA was its amnesty provisions, under which legal status was made available to illegal aliens already in the U.S. if they had been here continuously and illegally since 1982. In the late 1980s and early 1990s 2.7 million formerly illegal immigrants were legalized under this provision. Many of these individuals have since been joined by their spouses and children -- "legalization dependents" under the 1990 Immigration Act (cf., the Ghonda story).
One condition of the amnesty was that the aliens were ineligible for welfare, food stamps, and other federal government benefits for five years after gaining legal status. Further, the act prohibited discrimination against legal aliens (cf., political economy).
As discussed by Espenshade et al., last year's welfare reform legislation extended the denial of access to welfare and other social services to all legal immigrants during their first five years in the U.S. Hence, whereas previously the key distinction had been between legal and illegal immigrants, current (welfare) legislation now denies benefits to both groups of immigrants. These provisions are based on arguments that generous welfare benefits induce immigration from poor nations, with adverse consequences in low-wage labor markets, and increase tax burdens to pay for services to immigrant families.
At present, the Immigration Act of 1990 is the governing piece of legislation. The act provides for an annual ceiling of 675,000 immigrants, including 480K to be family-sponsored, 140K employment-based (all but 10K of which are for professionals and skilled labor), and 55K as "diversity immigrants" (cf., green card lottery). Hence, 71 percent of the authorized ceiling is family-oriented, 21 percent employment-oriented, and 8 percent diversity-oriented.
However, "legalization dependents" and other immediate relatives of U.S. citizens (spouses, children under age 21, parents of citizens age 21 or older, and adopted orphans) are not counted against the ceiling. This has two interrelated consequences: first, recent legal immigration has been well in excess of 675K; and second, distinctly less than 20 percent of recent legal immigration has been for employment purposes.
Immigration flows and composition
The effects of the policy changes that took place in 1965 and subsequently may be readily seen in Table III.C.4. During the 1950s the U.S. received about a quarter million immigrants per year, with nearly 60 percent of them coming from Europe. The flow of legal immigrants has risen steadily over time to well over a million per year (more than three quarters of a million per year excluding IRCA legalizations) during the 1990s. At the same time, the composition of that flow has been altered quite substantially.
Europe contributed only 10 percent of immigrants during the 1980s after having sent a substantial majority in the 1950s. The most spectacular shift over time offsetting the decline in European immigration has been in immigration from Asia. There was a six-fold increase in the Asian share of the total number of immigrants from the 1950s to the 1970s and 1980s, and in the absolute number of immigrants the increase was by 20-fold and more from the 1950s to the early 1990s. North America has become a distinctly more important source of immigrants to the U.S., and this flow -- which in the 1950s included many from Canada -- has increasingly come from Mexico and the Caribbean.
With the end of the IRCA legalizations, the dominance of Mexico as a source of immigrants appears to be subsiding (it accounted for more than one third of immigrants during 1991-93, outnumbering Asian immigration). The relative contributions of Asia and the Caribbean are rising, returning to their relatively high levels of the 1980s and 1970s, respectively.
Policy issues
There are at least three major policy issues pertaining to immigration that have received considerable attention of late. First is the question of how the "new" immigrants perform in and adapt to the labor market. Second, as noted above, is the question of what impact immigrants have on the U.S. labor market, and most notably the low-wage labor market (but cf., recent efforts by U.S. mathematicians to restrict immigration of Ph.D. mathematicians). These two labor market issues have long been studied by economists.
Finally, also as mentioned above, there is substantial concern about the fiscal effects of immigration. Immigrants pay taxes and at the same time receive various benefits -- the question here is basically, are they a net drain on public expenditures? Further, since immigrants tend to concentrate heavily in particular states -- most notably California, with close to a quarter of recent immigrants, and New York, with 17-18 percent -- there is concern that the incidence of fiscal drain will be especially heavy in immigrant-receiving jurisdictions.
With respect to the first issue, the performance of immigrants in the U.S. labor market, Borjas describes the "received wisdom of the early 1980s": immigrants started out at a disadvantage relative to Americans born in the USA (given their age and education, etc.), but over time the gap narrowed. Studies estimated "catch-up" times on the order of 15 years (with this catching up seen as reflecting the dynamism and upward mobility -- i.e., selectivity -- of immigrants), with some variation by country of origin (cf., culture-specific skills).
For example, Chiswick's analyses of 1980 census data on the earnings of foreign-born men found that relative to those born in the United Kingdom, those from Canada, Ireland, and elsewhere in Europe had lower earnings by about 5-15 percent, other things equal (i.e., education, experience, period of immigration, place of residence, marital status). The differences for men from elsewhere in North and South America were approximately 30 percent, while those for men from Asia ranged from 15-30 percent. Further, there was a clear tendency across groups for relative earnings to improve as time spent in the U.S. increased.
Borjas clearly believes that this received wisdom is already out of date. That is, with the increased emphasis on family reunification over occupational skills since 1965, the gap in education and hence in earnings between immigrants and natives has been growing. Borjas goes so far as to argue that "Because the newest immigrant waves start out at such an economic disadvantage, and because the rate of economic assimilation is not very rapid, the earnings of the newest arrivals may never reach parity with the earnings of natives."
It is not evident that Borjas's pessimism concerning parity of earnings is entirely well-founded (cf., Chiswick view). However, the issue of assimilation is raised by the growth of ethnic enclaves (e.g., little Saigon) and by the key role now of command of the English language as an influence on immigrants' earnings.
In any case, the notion of the U.S. as a "melting pot" for immigrants clearly seems less relevant now than in the past. Whether or not the current situation will result in lots of poor immigrants simply happy to be better off than in their home countries, and transmitting their poverty to their children (as Borjas suggests), remains to be seen.
The second issue that has received considerable attention is the impact of immigrants on the U.S. labor market. Pursuant to provisions of the Immigration Act of 1990, the U.S. Commission on Immigration Reform (created by the Act) has commissioned a study by the National Research Council on the impact of immigration on national and regional labor markets. [The study, due out next month, will also examine the effects of legal and illegal immigration on the future size, composition, and geographic distribution of the resident population, and the fiscal impacts of immigration on federal, state, and local governments.]
With the legalization of formerly illegal immigrants under the IRCA amnesty provisions as well as the decline in relative educational attainment of immigrants entering under family reunification provisions, the concern here has focused on the role that immigrants may have had on low-wage labor markets in particular. Borjas describes the "received wisdom of the early 1980s" as holding that immigrants did no harm to native employment opportunities. He argues that now, however, the substantial volume of immigration of lower-skilled workers has harmed the economic opportunities and relative wages of less-skilled and lower-educated native workers.
That is, as noted in the article by Wilson and Jaynes, because unskilled jobs require little or no training, there is a presumption of a high degree of substitutability between immigrants and unskilled native workers. Conversely, low substitutability between most immigrants and skilled native workers means that the latter group will not be harmed by immigration. Indeed, as Borjas notes, they benefit as consumers from the low wages of unskilled immigrants, and the increased demand for goods and services resulting from immigration serves to benefit skilled workers as producers as well.
Contrary to the rather negative assessment of Borjas concerning the labor market impact of immigrants, Wilson and Jaynes present a more complex picture. They emphasize the importance of geographic differences, which certainly seems plausible given the concentration of immigrants in some areas and their absence from others.
Wilson and Jaynes argue that estimates of the effects of immigration should be disaggregated not only by region or location, but also by ethnic group (reflecting the existence in some cases of ethnic group networks in certain labor markets) and by industry. After doing analyses along these lines (and also exploring the impact of immigration on internal migration of native workers), they conclude that "the interactions among immigration and internal migration, joblessness, and wages of native and foreign-born workers present no simple or straighforward story, when one begins to tease apart the broad categories of 'immigrant' and 'native-born' into more precise regional, occupational, and industrial categories."
The final policy issue to consider is the fiscal impact of immigration. Again, Borjas notes that the earlier received wisdom (that immigrants were less likely to receive welfare assistance than natives) is no longer true. He contends that 21 percent of immigrant households now participate in means-tested social assistance programs, compared to only 14 percent of native households. Hence, he concludes that "immigration may create a substantial fiscal burden on the most-affected localities and states."
However, Frank Bean and colleagues from the University of Texas have argued that comparisons of receipt of welfare among immigrants and natives may fail to accurately capture meaningful differences. There are questions about the reliability of data used (especially for Census data, which are the most comprehensive but do not identify directly the program in which public assistance recipients are enrolled), and about the appropriate unit of analysis (household, family, or individual -- where these may differ because of horizontally or vertically integrated households).
Further, Bean et al. argue that the magnitude of differences in welfare receipt between immigrants and natives depends upon the kind of immigrant. In this regard, they find that the rise in welfare recipiency among immigrants during the 1980s stems in large part from growth in the number of Asian refugees (cf., Vietnamese and others); and they argue that to the extent that this represents a problem it is a problem of refugee policy rather than of overall immigration policy. Likewise, they argue that growth in the population of Mexican and Central American immigrants, via IRCA legalizations, has had some consequences that may have greater implications for border control policies than for determining the overall levels of legal immigration.
The big immigration policy questions
The three policy issues just discussed will get additional exposure next month when the National Research Council report on demographic and economic consequences of immigration is issued. However, it seems clear that this report will not end the debates about the desirable level of and criteria for immigration visas (cf., Negative Population Growth ad from last November's Atlantic Monthly).
These two aspects are the big questions for immigration policy. There is currently considerable sentiment for reducing the overall level of immigration from their present high levels. Further, Borjas makes a case for swinging the pendulum back in favor of putting more weight on economic considerations and less weight on humanitarian (family reunification) considerations in the criteria for issuing immigrant visas.
He emphasizes the fact that among native Americans there are both winners and losers from different immigration policies: "this fact goes a long way toward explaining why the debate over how many and what kinds of immigrants to admit is best viewed as a tug-of-war between those who gain from immigration and those who lose from it."