During the 20th century the U.S. foreign-born population became increasinglyconcentrated in just a few states. The post-1965 increases in immigration levels and shifts inorigins reinforced the trend toward geographic concentration with a specific tilt towardCalifornia, especially, but also Texas and Florida. By 1990, more than one-third of allimmigrants lived in California and more than 70 percent in the top six states. There was everyreason to expect this growing concentration to continue since more than one-third of all newimmigrants were headed for California and the state was receiving more foreign-born migrantsfrom other states than it lost to them. These patterns continued through the early 1990s.
Beginning in the 1990s a dramatic shift occurred. The share of the immigrant populationliving in the six states with the largest immigrant populations declined for the first time indecades, dropping from more than 73 percent to below 70 percent in 1999. California's shareslipped from 34 percent to 30 percent as its foreign-born population stopped growing during the latter half of the 1990s.
This shift occurred as a result of a large drop in the share of immigrantssettling in California and a net flow out of the state among the settled immigrant population, anoutflow made up primarily of low-skilled Mexican-born immigrants.
Meanwhile, the immigrant population overall has continued to grow, but manyimmigrants are settling in new places. Our analysis shows that the share of growth andpopulation lost by California (and the other top six states) was almost entirely shifted to a groupof 19 "new growth" states. These states are in a swath reaching all the way across the entirecountry and are generally experiencing healthy levels of job and population growth overall.
The shifts in settlement patterns of the foreign-born population fit well within theimplications of migration theories. Prior to 1990, a very high percentage of immigrants settled ina few areas where previous immigrants had settled. Then, in the 1990s (and perhaps earlier),some small shifts occurred. A new immigrant (or perhaps a migrant from California) ended upin a nontraditional area. There were few immigrants in these areas, but the settlers found jobs,economic opportunity, and hospitable living conditions. Over time, friends, relatives, andcountrymen of the original immigrants moved to the new areas from California and othertraditional settlement areas. Then, new immigrant streams developed, bypassing the traditionalsettlement states and going directly to the new areas (see Massey et al. 1993). This stereotypicaltale of migration fits very well with the pattern shown by our datanet internal migration intothe new growth areas in the early 1990s followed by a shift in destinations of new immigrants inthe late 1990s.
This dispersal of the immigrant population to new settlement areas is occurring at thesame time that federal welfare reform devolved to the states new policymaking authority andfiscal responsibility for immigrants. These coinciding trends of dispersal and devolution raise
the question of whether immigrants are moving to states providing the most generous benefits tothem. Our analysis indicates that this is, indeed, not the case as immigrants appear to be movingmainly for higher-paying jobs in areas of growing employment.
The new dispersal trends outlined here raise a number of questions that still need to beexamined. These include:
- How will immigrants integrate into next destination communities? Will patterns of
integration differ substantially from patterns of integration seen in the moretraditional destinations?
- How will states respond to growing immigration? The new growth states have higher
shares of post-1996 immigrants than other states, yet they have made among the leastgenerous choices with regard to eligibility for benefits. Over time the number andshare of these immigrants will only grow, raising the question of how they will fare instates with comparatively weak safety nets and little experience integratingimmigrants.
- What impact will immigration growth have on local labor markets, schools, the health
care system, and the nonprofit structure, especially in light of the recent slowing ofthe country's economic expansion?
These new settlement patterns have not been fully incorporated into "commonknowledge" or demographic and social databases. The critical nature of immigration forunderstanding population change is underscored by the early results from Census 2000 whichfound significantly more people than expected overall and approximately 10 percent moreHispanics and Asians than anticipated," groups especially affected by the new immigrationpatterns (Passel 2001). The Hispanic population grew much faster between the 1990 and 2000Censuses in the new immigration states than elsewhere-127 percent from 2.3 million to5.3 million versus 43 percent in California, 48 percent in the top six immigration states, 71 percent in the traditional immigration states, and only 52 percent in the remaining states. Avery similar pattern shows up for Asians with 111 percent growth in the new immigration states_versus 52 percent in California, 64 percent in the top six immigration states, 93 percent in thetraditional immigration states, and 52 percent in the remaining states.
Further, initial analysis of the state data shows at least two results that can be traced toimmigration patterns. First,;he census count for the total population of California was somewhatlower than the predictions based on population estimates. This result is particularly notable in acontext where the overall count is 5 million or more above expectations and the national totalsfor the state's principal minorities are well above expectations. Second, the census counts forHispanics and Asians are disproportionately higher in the new growth states than in otherstates.18Put simply, this pattern of systematically higher-than-expected counts for Hispanics andAsians overall, but especially in the new growth states, combined with a low count for Californiaarose because the Census Bureau failed to fully measure immigration and relied on data from the1980s to estimate where the new immigrants were settling rather than data reflecting the patternswe describe in this report. Analysts who relied on the Census Bureau's population estimates andINS data on immigration could easily have missed the full impact of the new settlement patterns(e.g., Frey and DeVol 2000).
Eventually, the data from Census 2000 will permit further investigation of theimmigration and settlement patterns described here. These new data should not only reinforce
our results, but may ultimately describe even greater shifts.° The census can providesignificantly more detail in terms of geographic specificity (i.e., metropolitan areas and cities inaddition to states) and the characteristics of immigrants and migrants to study settlementpatterns. Finally, settlement and integration patterns, particularly for local areas, can beaddressed in much more detail with the census data than is possible with the