The geographic diversification of American immigration

A salient characteristic of immigration throughout the world is its geographic concentration. Immigrants tend not to disperse randomly throughout destination nations, but to move disproportionately to places where people of the same nationality have already settled. To a large extent, this selective channeling of immigrants to specific destination areas reflects the influence of migrant networks (Massey 1985). Because international migration is costly in both monetary and psychic terms, migrants display a strong tendency to draw upon social ties they have with current or former migrants in order to reduce the costs and risks of international movement (Massey et al. 1998). For someone contemplating a trip to the United States without documents, knowing someone with prior migratory experience can dramatically increase the odds of crossing the border successfully, finding food and lodging, and securing a good, steady job (Massey et al. 1987). Migrant networks have long functioned to channel migrants to specific places of reception. Their operation was noted early in the twentieth century by William I. Thomas and Florian Znaniecki (1927) and they have been rediscovered in the post-1965 era (Tilly and Brown 1967; Choldin 1973; MacDonald and MacDonald 1974). In both eras, migrants drew upon networks to migrate in response to structural transformations arising from the creation and extension of markets in the course of economic development (Massey and Taylor 2004). The recent period of globalization is no exception. Unlike the earlier era of globalization, however, the present one is characterized by the emergence of a select set of "global cities" that serve as centers of command and control for world markets (Sassen 1991). These urban centers house corporations that manage global capital flows and coordinate the international division of labor, attracting skilled and unskilled workers from throughout the world. As a result, a very few cities (New York, Los Angeles, Chicago, Houston, Miami) and a very few states (New York, New Jersey, California, Illinois, Texas, and Florida) receive a disproportionate share of America's immigrants, much larger than the shares received by leading cities during the classic period of immigration early in the twentieth century. Whereas, between 1901 and 1930, 36 percent of all immigrants went to the five most important urban destinations and 54 percent when to the top five destination states, from 1971 to 1993 nearly half of all immigrants (48 percent) went to the top five urban areas and 78 percent went to the five most important states, leading Douglas S. Massey (1995) to conclude in 1995 that geographic concentration was a quintessential and distinctive feature of the new immigration compared with the old. During the 1990s, however, something dramatic happened-there was a marked shift of immigrants away from global cities and the states or regions where they are located toward new places of destination throughout the United States. Not only did immigrants flow into states that hadn't received immigrants in any number since the 1920s such as Pennsylvania (see Shutika 2005); in many cases they entered states that had never before experienced any significant immigration of any sort, such as North Carolina (Bailey 2005; Griffith 2005), Georgia (Hansen 2005; Hernández-León and Zúñiga 2005), Virginia (Bump 2005), Arkansas (Schoenholtz 2005), and Louisiana (Donato, Stainback, and Bankston 2005), Kentucky (Rich and Miranda 2005), and Delaware, Maryland, and Virginia (Dunn, Aragonés, and Shivers 2005). To date, most of the scholarly work has focused on the geographic deconcentration of Latin American immigration (Zúñiga and Hernández- León 2005), although some investigators have considered the integration of Asians and other non-Hispanic groups in new areas of reception (Bump, Lowell, and Pettersen 2005; Fennelly 2005). For the most part, however, the literature consists of case studies of interactions between Latin Americans and non-Hispanic black and white natives, typically within places heretofore bereft of a significant Hispanic presence (Bailey 2005; Bump 2005; Solórzano 2005; Schoenholtz 2005; Gouveia, Carranza, and Cogua 2005; Shutika 2005; Grey and Woodrick 2005; Dunn, Aragonés, and Shivers 2005; Rich and Miranda 2005). Studies suggest the geographic transformation has been particularly dramatic in the case of Mexicans, the nation's largest immigrant group (Massey, Durand, and Malone 2002). Prior to the Great Depression, Mexican immigrants concentrated heavily in the Southwest, with 86 percent of all migrants going to just three states: Texas, California, and Arizona. Some 55 percent of Mexican immigrants between 1900 and 1929 went to Texas alone (Durand, Parrado, and Massey 1999). After the Second World War, California surpassed Texas as the leading destination and Illinois eclipsed Arizona to yield a new rank ordering of destinations. By 1970, 53 percent of Mexican immigrants in the United States lived in California, 27 percent resided in Texas, 6 percent were in Illinois, and 5 percent lived in Arizona (Durand, Massey, and Capoferro 2005). As the new immigration gathered steam after the implementation of the 1965 amendments to the Immigration and Nationality Act, Mexican immigration grew even more concentrated geographically, focusing increasingly on California. Among those Mexicans who arrived in the United States between 1975 and 1980, for example, 59 percent went to California, 21 percent went to Texas, and 9 percent went to Illinois, together accounting for nine of ten newly arriving Mexicans (Durand, Massey, and Capoferro 2005). This pattern of regional concentration persisted through the 1980s. Among Mexican immigrants who arrived in 1985-1990, nearly two thirds (63 percent) went to California, with another 20 percent going either to Illinois or Texas (Durand, Massey, and Capoferro 2005). Sometime after 1990, however, there was a dramatic shift in migration patterns. California fell sharply in popularity as a destination for Mexican immigrants, and though the decline was less severe for Texas and Illinois, they too fell in importance. Among Mexicans who arrived in the United States between 1995 and 2000, only 28 percent went to California, 15 percent to Texas, and 6 percent to Illinois. Rather than accounting for 90 percent of the inflow, these three states suddenly accounted for less than half (Durand, Massey, and Capoferro 2005). Despite the research done to date, we still lack a clear description of the changing geography of American immigration. Jorge Durand, Emilio A. Parrado, and Massey (1999) and Durand, Massey, and Chiara Capoferro (2005) focused exclusively on the Mexican case. Micah N. Bump, B. Lindsay Lowell, and Silje Pettersen (2005) only examined the period from 1990 to 2000 and only contrasted Latin American and Asian immigrants in the year 2000. Here we focus on foreigners who entered the United States in the five years prior to the census or survey to measure trends in immigrant destinations during the late 1970s, late 1980s, the late 1990s, and during the first five years of the new century, comparing patterns for Mexicans, other Latin Americans, Asians, and immigrants who were neither Asian nor Latino. We seek to understand the degree to which geographic diversification characterizes other national-origin groups besides Mexicans, whether the beginnings of the geographic transformation can be detected prior to 1990, and whether the pattern of diversification is being sustained into the new century.