Donate
Independent, objective, nonpartisan research

R 709SBR

Authors

R 709SBR

Tagged with:

Publication PDFs

Database

This is the content currently stored in the post and postmeta tables.

View live version

object(Timber\Post)#3711 (44) { ["ImageClass"]=> string(12) "Timber\Image" ["PostClass"]=> string(11) "Timber\Post" ["TermClass"]=> string(11) "Timber\Term" ["object_type"]=> string(4) "post" ["custom"]=> array(5) { ["_wp_attached_file"]=> string(12) "R_709SBR.pdf" ["wpmf_size"]=> string(6) "537102" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(89212) "New Patterns of Immigrant Settlement in California Sarah Bohn with research support from Eric Schiff Supported with funding from The Ford Foundation and from t he Research Foundation of The City U niversity of New York July 2009 The Public Policy Institute of California is dedicated to informing and improving public policy in California through independent, objective, nonpartisan research on major economic, social, and political issues. The institute’s goal is to raise public awareness and to give elected representatives and other decisionmakers a more informed basis for developing policies and programs. The institute’s research focuses on the underlying forces shaping California's future, cutting across a wide range of public policy concerns, including economic development, education, environment and resources, governance, population, public finance, and social and health policy. PPIC is a private operating foundation. It does not take or support positions on any ballot measures or on any local, state, or federal legislation, nor does it endorse, support, or oppose any political parties or candidates for public office. PPIC was established in 1994 with an endowment from William R. Hewlett. Mark Baldassare is President and Chief Executive Officer of PPIC. Walter B. Hewlett is Chair of the Board of Directors. Copyright © 2009 by Public Policy Institute of California All rights reserved San Francisco, CA Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source and the above copyright notice is included. Research publications reflect the views of the authors and do not necessarily reflect the views of the staff, officers, or Board of Directors of the Public Policy Institute of California. 3 Contents Summary 4 Acknowledgments 5 Introduction 6 CALIFORNIA’S POPULARITY DECLINE 13 CHANGES IN THE COMPOSITION OF CALIFORNIA ’S NEW IMMIGRANT ARRIVALS 19 CHOICES WITHIN CALIFORNIA 22 HAVE NEW IMMIGRANTS’ PREFERENCES CHANGED? 27 Have Latino Immigrant Preferences Changed? 28 Have Asian Immigrants’ Preferences Changed? 30 Conclusion 35 References 36 About the Author 38 4 Summary California has been a crucible for immigrant -related issues in the United States for decades due to the overwhelming share of U.S. immigrants who choose to live in the state . However, in the late 1990s, the popularity of California among immigrants began to decline for the first time in nearly 100 years. As fewer immigrants locate to California and to other traditionally immigrant -rich areas, more are choosing to live in states and cities with little history of immigration. One consequence of this demographic change is a rise in immigration -focused local legislation in those new settlement areas. In the first half of 2008 alone, more than 1,200 state bills related to immigration were proposed across the country . Immigration is no longer an issue limited to a n isolated handful of states or a handful of cities; it affects many areas across the country. In light of this demographic shift and associated policy responses, and to provide context for the immigration debates being waged nationwide and locally, this study examines immigration in detail using U.S. Census data from 1990 –2007 . We find that the decline in California’s share of the nation’s immigrant population is driven partly by out -migration of established immigrants to other states, but mostly by the settlement of new immigrant arrivals in to different states. California has experienced a net out- migration of both established immigrants and native-born persons to other states, but the flow of established immigrants is relatively small. In contrast, California’s share of new immigrant arrivals to the United States has fallen sharply, from 35 percent of new arrivals in the late 1 980s to only 19 percent in 2004 –2007. Although new immigrants to the United States and to California are not markedly different from their predecessors, it appears that some of the socioeconomic considerations related to their migration choices have changed . Social factors such as residence near co -ethnics explain a large portion of immigrant concentration in California, but these have waned in importance . At the same time, economic factors have remained consistent in explaining where immigrants tend to locate. For those concerned with the integration of immigrants, the decline in clustering of immigrants along social dimensions may be good news . The relevance of economic opportunity in immigrants’ location decisions ma y also bode well for their economic integration . On the other hand, in areas with few immigrants and little experience incorporating immigrants into social and economic life, the settlement of immigrants away from co -ethnics may lead to increasing isolation for these groups and new challenges for the communities they settle in . All technical appendices to this paper ar e available on the PPIC website: http://www.ppic.org/content/pubs/other/709SBR_appendix.pdf 5 Acknowledgments Eric Schiff provided excellent research support for this project. The author acknowledges the helpful reviews of Hans Johnson, Laura Hill, Audrey Singer, and Randy Capps, as well as input from Dylan Conger, Hector Cordero -Guzman, Jason Fitchener , seminar participants at the Association for Public Policy Analysis and Management 2008 Fall Conference , and at the American Economic Association 2009 Annual Meeting. 6 Introduction By 2007, an estimated 38 million immigrants resided in the United States. About 26 pe rcent of those lived in California . For 90 years, the popularity of California as a destination for immigrants steadily increased . But in the 1990s, for the first time since the early 1900s, Cali - fornia’s draw for immigrants began to wane. The rapid immigrant growth over most of the state’s history has helped shape its demo - graphic and economic makeup, and has fueled fierce debate, as epitomized by Proposition 187 in 1994 and Proposition 227 in 1998. 1 Figure 1 . Percent of U.S. Immigrants Living in California, 1850 –2007 The marked turnaround in growth in the 1990s raises a host of questions. Among them are: why did the decline occur and what does it mean for California’s future? SOURCE: Passel and Zimmerman (2001), US Census Bureau (1999), author's calculations from Census data. California has a unique history with immigration, in terms of the number of immigrants who choose to live in the state and because of its consequent policy experience with immigrant- related issues. However, the decline in popularity of the state as an immigrant destination is not unique ; other leading immigrant destination states have experienced this decline as well . As Map 1 shows, the states with the largest concentration of immigrants in 2000 ex perienced some of the smallest increases in the number of immigrants from 2000–2007. 2 1 Proposition 187 was a broad -ranging measure dealing with immigrants in California that empowered all law enforcement agents to enforce federal immigration law and imposed restrictions on public benefits for immigrants. Proposition 227 eliminated bilingual education. California, New York, Illinois, Texas, Florida , and New Jersey have the largest immigrant 2 This map counts only immigrants of working age, 18 –64. See Singer (2004) for a similar map with all immigrants included for 1990-2000. 1.03.53.84.44.03.64.35.47.68.0 10.2 13.8 18.3 25.4 32.7 28.5 26.4 0 5 10 15 20 25 30 35 40 1850186018701880189019001910192019301940195019601970198019902000 Percent Ye a r P e rc e nt in CA (va lue s shown) P e rc e nt in CA of working a ge P e rc e nt of US P opula tion in CA 7 populations, but higher growth is found in n ew immigrant destination states. The same is true for immigration during the 1990– 2000 period. Map 1. Percent Change in Number of Immigrants , 2000 –2007 SOURCE: Authors’ calculations from IPUMS Decennial Census data. NOTE: Includes only immigrants age d 18-64 . Shading represents quintiles of the percentage change. The dispersion of immigrants across the United States is a relatively new phenomenon, and has been noted by many researchers. In 1990, 74 percent of all immigrants of working age in the United States lived in six states. Californi a alone was home to nearly 33 percent of them. This concentration had increased consistently over the previous 100 years of immigration history in the United States. 3 Throughout this report, an immigrant is def ined as a person born outside the United States and its territories and either a naturalized American or non -citizen. The documentation status of immigrants is a particularly heated aspect of the debate about immigration nationally and locally . T here are n o comprehensive data that allow researchers to identify the documenta - tion status of immigrants at the individual level, so this study will consider all immigrants For the first time, in the late 1990s, the percent of immigrants living in the top six states declined , falling to 69 percent in 2000 and to 66 percent in 2007. Although these decreases seem small against the backdrop of nearly 100 years of increasing concentration , the trend reversal is striking. California continues to be home to the largest number of immi -grants and has the highest concentration of immigrants (the ratio of immigrants to total popula -tion), but California rank ed 40th in the nation in percentage change in the number of immi -grants from 1990 to 2000 and 43rd from 2000 to 2007 . Between 1980 and 1990, California’s working -age immigrant population grew 9 .5 percent per year . This growth was down to 4 .4 percent per year between 19 90 and 2000 and to 2 percent per year between 2000 and 2007. 3 Passel and Zimmerman (2001). 8 regardless of legal status. I t has been argued that state and local policies aimed at illega l immigrants have spillover effects on all immigrants. 4 This demographic change in the United States overall raises important questions about how we understand the movement of population and how that movement contributes to the socioeconomic characterist ics of different areas. The academic literature on migration is well developed, but it has yet to fully analyze this re cent pattern. For California in particular, the decline in popularity of the state as an immigrant destination is not well understood. Have newly arrived immigrants simply decided to live in new places? Have established immigrants migrated away from California? And if so, are the same factors driving immigrants to new places affecting the location decisions of native -born California residents? These questions can be asked not only of immigrant settlement patterns between Cali - fornia and other states, but also of patterns within California . Within the state, Los Angeles County dominate s in number and concentration of immigrants. With nearly 2.3 million immi- grants in 1990 and 2.9 million in 2007, Los Angeles County has more than triple the number of immigrants of any other county . However, the county experienced very little growth in immi - grant population over this same period, especially compared to other counties . Between 1990 and 2006, the number of immigrants in Los Angeles County increased an average of just 1.9 percent per year, compared to rates as high as 12.6 percent annually in Riverside County and 10.5 percent in Kern County. Map 2 show s that the California counties with the largest immigrant populations experienced relatively low growth in immigrant population between 1990– 2000 and 2000 –2007. 5 Map 2 . Percent Change in Number of Immigrants, 2000–2007 SOURCE: Authors’ calculations from IPUMS Decennial Census data. NOTE: Includes only immigrants age d 18–64. 4 Singer et al (2008). 5 Census data allow us to identify 42 county groups in California. Small counties such as Sierra, Plumas and Nevada are grouped together because there is not enough information to accurately break out these counties individually. 9 Looking specifically at metropolitan areas within the state, we see that the Los Angeles Metropolitan Statistical Area (MSA) 6 Table 1 . Changes in immigrant population in California MSAs, 1990–2007 dominates all other state MSAs in terms of immigrant population . The Los Angeles MSA was home to 2.7 million immigrants of working age in 1990 and 3.7 million in 2007; this is about three times the size of the immigrant population in any other MSA . The San Francisco -Oakland MSA is the second la rgest, with about 629,000 in 1990 and 1.1 million in 2007. Number of immigrants Percent of MSA population Percent change MSA name 1990 2000 2007 1990 2000 2007 1990 –2000 2000 –2007 Bakersfield 48,246 84,365 129,227 15.7 23.7 28.7 74.9 53.2 Chico 8,031 10,998 13,774 7.6 9.3 10.0 36.9 25.2 Fresno 84,787 151,247 191,570 22.1 29.1 31.3 78.4 26.7 Los Angeles-Long Beach 2,693,286 3,493,571 3,654,601 37.7 46.0 45.6 29.7 4.6 Merced 25,458 38,174 49,491 25.8 33.0 34.7 49.9 29.6 Modesto 39,984 65,045 80,409 18.6 25.0 25.8 62.7 23.6 Redding 2,530 4,629 4,694 3.0 5.0 4.3 83.0 1.4 Riverside-San Bernardino 271,754 490,946 761,629 18.3 26.8 31.3 80.7 55.1 Sacramento 101,296 170,761 245,361 11.1 17.4 21.0 68.6 43.7 Salinas-Seaside-Monterey 57,542 62,833 64,319 28.3 40.0 39.1 9.2 2.4 San Diego 325,728 473,439 521,239 21.3 27.9 28.8 45.3 10.1 San Francisco -Oakland-Vallejo 628,890 967,173 1,078,199 23.9 32.6 35.4 53.8 11.5 San Jose 273,524 470,193 530,580 27.8 43.1 47.7 71.9 12.8 Santa Barbara-Santa Maria-Lompoc 47,123 66,564 68,661 20.6 27.9 29.0 41.3 3.2 Santa Cruz 25,257 39,623 37,710 17.6 24.1 23.2 56.9 -4.8 Santa Rosa-Petaluma 25,530 48,578 61,592 10.8 17.3 20.9 90.3 26.8 Stockton 54,521 86,697 130,597 19.8 27.3 32.8 59.0 50.6 Ventura-Oxnard-Simi Valley 87,987 125,388 150,836 21.3 27.6 30.6 42.5 20.3 Visalia- Tulare-Porterville 40,252 64,482 82,602 23.6 31.4 33.9 60.2 28.1 Yuba City 10,825 18,125 21,126 15.3 22.9 21.4 67.4 16.6 California MSAs 4,852,551 6,932,831 7,878,217 27.5% 35.5% 36.8% 42.9% 13.6% SOUR CE: Author ’s calculations from IPUMS Decennial Census and American Community Survey data. N OTE : Includes only immigrants age d 18- 64. San Diego, Riverside -San Bernardino, and San Jose MSAs are the next largest, but have only roughly half as many immigrants as the San Francisco-Oakland MSA. D espite the dominance of a few MSAs in terms of immigrant population, the MSAs that experienced the largest changes in immigrant population over the period were not those with a history o f attracting many immigrants (Table 1) . 6 There are 23 Metropolitan Statistical Areas (MSA) defined for California, which identify roughly the various centers of population and economic activity in the state. An MSA is defined around a population center and may be comp rised of single or multiple counties. Although they do not cover all of the land of the state, the population in MSAs in California comprises at least 95% of the total state population over 1990 -2006. For this reason, the trends for California’s MSAs mimic the trends described above for California counties. 10 The changing settlement patterns of immigrants have occurred at the same time as some immigrant -related developments in policymaking at the state and local level. In 2007, 1,562 immigration -related bills were introduced by state policymakers across the United States, about three times the number introduced in 2006. 240 of the bills introduced in 2007 were enacted . In the first half of 2008 alone, 1,267 bills were introduced and at least 175 signed into law. 7 Many state and local governments have taken policy action in response to the down - loading of immigran t-related issues and costs from the federal to the state and local level. This new legislation has covered a broad range of issues, including employment eligibility , human tr afficking, public benefits , and driving licens es. There is a lot of variation in the nature of these laws and ordinances: some are extremely restrictive while others are more accommodating . For example, Oklahoma’s HB 1804 makes it a felony to harbor or sh elter illegal immigrants and requires state and local law enforcement to enforce federal immigration laws. In contrast, California’s AB 976 prohibits landlords from asking about, or taking any action based on, a tenant’s immigration status. At the local le vel, it is likely that thousands of ordinances were proposed , although we are not aware of a source that has collected comprehensive information on them. 8 The federal government controls the number of immigrants legally allowed to enter the country each year, and to some extent controls the number of illegal immigrants th rough border control measures, arrests , and deportation s. State and local governments, however, have no direct power to regulate the number of immigrants who choose to settle within their borders. Thus, both the benefits and the costs of changes in immigrant population s accrue to local areas, but are largely out of the control of local governments . In the absence of federal immigration reform to assist local areas that receive large immigrant inflows, they have been left to deal with many of the challenges and costs on their own . This has lead to frustration at the local level, and a few lawsuits against the federal government. The policy levers that state and local governments can exercise range from those that deflect immigrants, such as enforcing zoning, licensing, an d housing codes, 9 and those that accommodate immigrants, such as establishing day labor sites and expanding bilingual education programs . This wide variety of responses is seen clearly in the range of state laws enacted in 2007 and 2008. 10 Previous studies have shown that changes in immigrant population, rather than its size, drive tensions at the local level. 11 7 Statistics in this paragraph are given in National Conference of State Legislatures January 31 and July 28 2008 reports on state legislation. The NCSL uses a comprehensive methodology for identifying all state legislation related to immigration. We are unaware of such a comprehensive methodology for local ordinances on the same subject. There are a variety of reasons. In non-gateway cities and states, the arrival of immigrants is a relatively new phenomenon . Since immigrants have hi storically settled in a very few number of places, there are a vast number of non -traditional immigrant destinations across the country that are ripe for tension. Indeed, the flurry of local legislation attempting to regulate immigration over the last few years reflects the potential for tension and conflict related to immigrant-driven demographic change. 8 Singer et al (2008), p. 157. 9 Light (2007). 10 While the effect and effectiveness of these policies with regard to immigrants is of great interest, not enough time has passed in order to fully identify the effects. We are planning future studies to carry out an evaluation of these policy changes. 11 Hopkins (2007), Singer et al (2008). 11 There is some evidence that areas with a long history of immigration become better at incorporating new immigrants. They may have established government services or community organizations offering assistance to immigrants. 12 In addition, ethnic enclaves have developed in many gateway cities that may both draw future immigrants and help them to become integrated, in particular by providing job opportunitie s. It may also be that over the long run, industries adapt to changes in population : the arrival of a large number of low skilled workers, for example, may induce a manufacturing firm to hire from the large pool of workers instead of investing in relatively more expensive machines that would otherwise replace workers. 13 Much of state regulation of im migration deals specifically with issues related to undocumented immigrants, including regulation of public benefits to illegal immigrants and restrictions on employment or housing . In California in 2007 –2008, 17 immigration -related laws were enacted and 1 5 resolutions were passed; nearly all of these laws and resolutions could be categorized as supportive of immigrants and immigration. Lastly, it may be that as fears about socioeconomic changes caused by the arrival of new immigrants are not realized, residents become less concerned . 14 The legislative surge relating to immigration issues begs the question of whether these new immigration patterns are related to the socioeconomic characteristics of immigrants or to the social and economic conditions in areas across the country. Immigrants have historically been clustered geographically, but they have also clustered around some socioeconomic characteristics . For example, immigrants in California are more likely to be Mexican, whereas immigrants in Florida are more likely to be Cuban. Similarly, t he San Francisco Bay a rea has tended to attract more highly educated immigrants, whereas the labor demand in other areas of the state attracts less educated immigrant workers. These dimensions of immigrant clustering affect local policy toward immigrants —for example, whether bilingual education programs or day labor centers are expanded —and they affect the demand for government-provided services and infrastructure . Thus, n ot only the fact that we experience new settlement trends is of interest, but also who is leaving, arriving, and staying becomes a concern . For example, if highly skilled new immigrants to the United States find California —or specific cities in California — less attractive, this trend may exacerbate the state’s problem of recruiting enough skilled workers. The same could be said of the six bills enacted in New York . The more accommodating nature of these bills may be evidence of these states’ long history of immigration . States with new immigrant growth tended to have a mix of legislative response to immigration issues. 15 Immigrant destination choice is also related to the success of immigrants themselves. We commonly frame the success of immigrants by their social and economic adaption to their destination of choice, commonly called assimilation. Historically, assimilation policies were often called “ Americanization” and largely referred to language learning, civic participation, naturalization, and acceptance of American cultural values (however defined) . B ut full immigrant assimilation also includes improving their economic outcomes with increased time Also, understanding the selectivity of migration may help policymakers estimate future funding needs for programs, services , and infrastructure. 12 Ramakrishnan and Lewis (2005). 13 Lewis (2005). 14 For additional detail on these bills, see Technical Appendix A. 15 Reed (2008). 12 spent in the United States and incorporation into social networks not related to becoming Americanized. Federal, state, and local policies aim to affect immigrant assimilation, for example, making bilingual education programs, job training, and migrant worker assistance services available . Local offices of i mmigrant affairs have long provided 16 We first examine the trend of immigrants locating to states other than California. Next, we look within the state at trends in the historically immigrant-rich are as and in new growth areas . Last , we provide an analysis of the underlying factors behind the change in immigrant location choice s and a discussion of the consequences. services to promote civic learning and to guide immigrants through the naturalization process. Recently, immigrant assimilation of this sort was a goal of former President George W. Bush’s Task Force on New Americans . Because the assimilation of immigrants is tied to their destination area and the reasons they choose to reside there, we explore the reasons behind immigrants’ changing settlement patterns. This may indicate prospects for immigrant assimilation and policies that might enhance assimilation. 16 California Senate Bill 1094 called for establishment of an Office of Immigrant Affairs. Other states and localities have similar offices with similar names. 13 California’s Popularity Decline Although California’s popularity as a destination among immigrants has declined since the late 1990s, the state is still home to the largest immigrant population in the country and that population has continued to grow . But t his growth in immigrant population in the state is much smaller than in the past and is much smaller compared to most states in the country . Table 2 shows the changes in immigrant population across states and groups of states durin g the period 1990 –2007. Table 2 . Changes in immigrant population, 1990– 2007 Number of i mmigrants Percent change Percent of U.S. immigrants State 1990 2000 2007 1990– 2000 2000– 2007 2000– 2007 on 10 year basis 1990 2000 2007 Total in U.S. 14,589,626 24,292,460 30,121,594 67 24 34 - - - Top Immigrant states - Total 10,778,377 16,770,500 19,897,845 56 19 27 73.9 69.0 66.1 California 4,933,152 7,101,428 8,083,580 44 14 20 33.8 29.2 26.8 New York 2,073,332 2,991,581 3,296,533 44 10 15 14.2 12.3 10.9 Texas 1,189,892 2,345,295 3,131,882 97 34 48 8.2 9.7 10.4 Florida 1,142,859 1,960,036 2,559,827 72 31 44 7.8 8.1 8.5 Illinois 720,573 1,214,660 1,429,508 69 18 25 4.9 5.0 4.7 New Jersey 718,569 1,157,500 1,396,515 61 21 29 4.9 4.8 4.6 Select high growth states - T otal 612,086 1,874,039 2,800,509 206 49 71 4.2 7.7 9.3 Arizona 199,473 507,084 784,769 154 55 78 1.4 2.1 2.6 Georgia 135,717 473,757 710,728 249 50 71 0.9 2.0 2.4 Colorado 107,796 288,427 391,970 168 36 51 0.7 1.2 1.3 North Carolina 88,311 348,962 505,699 295 45 64 0.6 1.4 1.7 Nevada 80,789 255,809 407,343 217 59 85 0.6 1.1 1.4 SOURCE : Author ’s calculations from IPUMS Decennial Census and American Community Survey data. N OTE : Includes only immigrants age d 18- 64. First, we note that the number of immigrants living in the United States increased by about 10 million from 1990 to 2000 and by about 6 million from 2000 to 2007 . A large fraction of the immigrant population lives in the historically popular destination states of California, New York, Texas, Florida, Illinois, and New Jersey. The United States and these six immigrant gateway states have seen increases in immigra nt population from 1990– 2007, but at a slower rate at the end of that period than at the beginning . In California, the immigrant population grew 44 percent from 1990 to 2000 but only 20 percent from 2000 to 2007 (on a 10-year basis) . Only eight states—Alaska, Hawaii, Louisiana, Maine, Montana, New York, Vermont, and West 14 Virginia —have lower immigrant population growth rates than California from 2000 to 2007. Since 1990, the share of immi grants choosing states with smaller immigrant populations and short histories of receiving immigrants has increased. In Arizona, Georgia, Colorado, North Carolina , and Nevada , immigrant populations grew at very high rates. 17 Changes in the immigrant population of California occur by various means. New arrival immigrants move into the state directly from other countries ; domestic (or internal) i mmigrants may also relocate from other states or to other states. In addition, some immigr ants may also leave the United States entirely. No nationally representative dataset is able to measure the outflow of immigrants from the United States accurately, so we do not consider the out- migration of immigra nts from the United States in this report. Although the growth rates in the se states declined from 1990 –2000 compared to 2000 –2007, as they did nationally, in four of them rates are nevertheless about double the rates of growth of the traditional states. T he share of immigrants choosing California fell by about 7 percent from 1990 to 2007 while the number choosing the new destination states rose about 5 percent over the same time frame. Indeed, t he decline in the percentage of new immigrants choosing the six gateway states is driven largely by the California’s decline . 18 Measuring Migration The 1990 and 2000 Census es and the American Community Surveys from 2005 to 2007 provide snapshot s of the population of states, counties, and MSAs . Looking at differences among these snapshots, we can estimate changes in population and demography. Because the Census asks the year of arrival for immigrants, we can estimate how much of the population change was due to new arrival immigrants. F rom this cross -sectional analysis we cannot estimate how much of the change was due to domest ic migration , but these datasets also allow study of migration at the individual level for a few distinct periods. In the 1990 and 2000 Census es, data were gathered on where respondents were living five years before. And in the American Community Surveys f or 2005 to 2007, data were gathered on where respondents were living one year before . From this, we can estimate the number of people that moved between states or MSAs over the 1985 –1990, 1995 –2000, and 2004 –2007 periods, allowing us to estimate the number of domestic in-migrants and out - migrants to a given area . Subtracting the number of out -migrants from the number of in- migrants gives the net internal migration of immigrants (or native-born) to the state or MSA. The components of net internal migration a re partly net figures already, as any intermediate moves within the period are not measured. Despite the drawbacks of measuring migration only within three- to five -year periods, the advantage of this Census data is the ability to break down migration sta tistics by individual 17 These new destination states are defined as being in the top 10 for growth and the top 25 for the number of immigrants. This definition gets around the problem of places with very few immigrants having very high growth rates because of the way growth rates are calculated rather than because there is a substantive change. 18 Various estimates suggest that up to 50 percent of immigrants to the United States eventually leave the country (Jasso and Rosenzweig (1982), Borjas and Bratsberg, (1996)). Some immigrants leave the country permanently, and some leave temporarily and return. The Census records the year of migration for immigrants, but does not ask whether this stay is the first or not. Our estimates do not consider out-migration and may be affected by immigrant s who have had multiple stays in the United States. Changes in death rates also affect the growth in California’s immigrant population, but given the relatively young average age of immigrants, changes in death rates are unlikely to affect overall estimates. 15 socioeconomic characteristics.19 We focus on persons of working age, because economic reasons tend to dominate migration decisions and because we are particularly interested in examining the effect of local labor market changes on migration patterns. (See T echnical Appendix A for background information on the determinants of location choice.) M ost analysis in this report is restricted to persons aged 18–64 who do not reside in institutions —the working age population. In parts of the analysis, we further restrict our attention to persons who report participation in the labor market (whether currently employed or unemployed). In particular, we are first interested in decomposing net internal migration into migration of native -born persons and migration of immigrants. As explained above, the stalling of growth of immigrant populations in historical immigrant destination states and MSAs may be caused by a decrease in new immigrant arrivals, an out -migration of previous immigrants, or both . In addition, changes in the concentration of immigrants, measured as a ratio of immigrants to total population, can also be caused by differential net internal migration of the native -born . So we proceed by first examining the number of new arrival immigrants and the net internal migration of the native- and foreign-born. We first examine whether the changes in immigrant population in California are driven by changes among newly arrived immigrants or among previous cohorts of immigrants. Table 3 shows that over these three periods, California’s growth in immigration population comes almost entirely from new arrivals. 20 Table 3 . Components of m igration in California Population change due to Change on a per-year basis California share of national new immigrant arrivals Net internal migration Net internal migration Years Net internal migration = Native born + Immigrant New immigrant arrivals Net internal migration = Native born + Immigrant New immigrant arrivals 1985–1990 120,714 81,333 39,381 942,795 24,143 16,267 7,876 188,559 34.9% 1995–2000 -407,162 -249,205 -157,957 906,935 -81,432 -49,841 -31,591 181,387 21.0 2004–2007 -388,374 -288,747 -99,627 698,836 -129,458 -96,249 -33,209 232,945 19.3 SOURCE : Author ’s calculations from IPUMS Decennial Census and American Community Survey data. N OTE : Includes immigrants age d 23– 64 as of 1990, 2000, or 2007, so as to count only persons of working age during the entire migration period. On an annual basis, the number of new immigrants to California was large and rela- tively steady during the periods 1985 –1990, 1995 –2000, and 2004 –2007. 21 19 We construct the 2004 –2007 migration period by linking one –year migration information across the three cross - sections of data. For this reason, we cannot track individual characteristics of 2004 –2007 migrants, only aggregate or group characteristics . However, California experienced an increase in the domestic out -migration of immigrants from 1995 to 2000 and 20 To view these migration trends relative to the state population, see T echnical Appendix Table A1. 21 For the comparison of the 2004– 2007 period to earlier periods, it is necessary to consider the population change on a per-year basis. Statistics from 1985– 1990 and 1995–2000 are divided by five and 2004– 2007 statistics are divided by three. Given the available data, this is the best comparison that we can do. However, the shorter migration period of 2004– 2007 may induce some bias. For example, we see a ri se in new immigrant arrivals in 2004–2007 on a per -year basis. It could be that the rate of new immigrants who arrived between 1995 –1998 was just as high but declined when the last two years of the period were measured. A comparison of new immigrants to Ca lifornia in the 2002–2007 and 1995– 2000 periods from the CPS suggests that indeed, the number of new immigrants fell when these full five-year migration periods are analyzed. Other limitations of the CPS data, however, prevent us from using it for the full analysis in this report. 16 from 2004 to 2007. During 1985–1990, more internal im migrants moved into the state than moved out, but during the latter two periods the opposite was true. Were it not for the increasing outflow of immigrants from California, the state would have seen a slightly larger growth in the immigrant population, because the inflow of new international immigrants remained roughly the same during the 1985–1990 and 1995 –2000 periods. Although on a per - year basis the inflow of new immigrants to California increase d from the 1995– 2000 period to the 2004– 2007 period , the percentage of all new immigrants who chose California continued to decline . In all periods, the number of newly arrived immigrants to California far outweigh ed the number of net migrants. So it is clear that the trends among new arrival immigrants drive the trends in overall immigrant population for the state. Comparing internal migration of immigrants to migration of native-born, we find that native -born net outmigration increased even more rapidly. 22 In addition, we find that California’s experience is unique among the immigrant gateway states , in that native-born outmigration mirrored that of immigrants, but did not accelerate as quickly . 23 In light of the declining popularity of California with new and previous immigrants, we next examine whether there is any evidence that the trend is driven by the choices of immigrants with certain socioeconomic characteristics. Who is it that California is no longer attracting? Are highly skilled immigrants choosing other states? Figure 2 shows the education distribution of California’s immigrants. In California and other traditional immig rant gateway states, the decline in new arrivals drive s most of the change in immigrant populations . In the new growth states, changes in immigrant populations are driven mostly by new arrivals as well, but are amplified by net increases due to internal mi gration of previous immigrants . In both group s of states, the trend among new immigrants is generally mirrored by the trend among domestic migrants. This marks a change from the recent past, where the flow of native-born domestic migrants tended to move in the opposite direction from the migration of immigrants. It also points to potential commonality in the reasons for migration between immigrants and the native-born . 22 This is a notable change from studies of migration over earlier periods that tend to find net domestic migrants move in the opposite direction from new immigrants (Frey and Liaw, 1998, and Bartel, 1989). However, Passel and Zim merman (2001) find a similar pattern for California during 1990–1995. 23 Technical Appendix Table A2 presents migration patterns for the six top immigrant states and the top new growth states. Technical Appendix Table A3 gives statistics for all states. 17 Figure 2. California’s Immigrant Education Distribution, 1990 –2007 SOURCE: Author ’s calculations from IPUMS Decennial Census and American Community Survey data . NOTE : Includes only immigrants age d 18–64 . The share of immigrants who had not completed high school fell from 39 percent in 1990 to 37 percent in 2000 and to 32 percent in 2007. Conversely, immigrants with a bachelor’s degree or higher increased from 17 percent in 1990 to 20 percent in 2000 and to 24 percent in 2007 . The shift in education distribution of immigrants can be explained primarily by changes in the education level of new immigrants. To show this, we decompose the components of migration in California along the education dimension. 18 Table 4 . Components of m igration in California by education Per year basis Net internal migration of New immigrant arrivals Percent of all new immigrants Years Education Net internal migration = Native born + Immigrant 1985–1990 < High school -4,562 -4,605 42 77,588 41 High school -6,272 -6,797 524 42,110 22 Some college 1,606 -520 2,127 32,179 17 College graduate 33,371 28,188 5,183 36,682 19 Total 24,143 16,267 7,876 188,559 1995–2000 < High school -25,806 -5,917 -19,889 61,315 34 High school -34,470 -24,071 -10,398 39,312 22 Some college -35,936 -31,016 -4,920 27,519 15 College graduate 14,779 11,163 3,615 53,242 29 Total -81,432 -49,841 -31,591 181,387 2004–2007 < High school -22,728 -6,147 -16,581 69,181 30 High school -43,509 -29,182 -14,327 51,427 22 Some college -47,928 -41,438 -6,490 30,149 13 College graduate -15,293 -19,482 4,189 82,188 35 Total -129,458 -96,249 -33,209 232,945 SOURCE : Author’s calculations from IPUMS Decennial Census and American Community Survey data . N OTES : Includes immigrants age d 23– 64 as of 1990, 2000, or 2007, so as to count only persons of working age over the entire migration period. Table 4 shows that there has been a net outflow of individuals with less than a high school diploma from California in all three periods . T his outflow is composed mostly of native- born p ersons from 1985 to 1990 but then largely of immigrants during the 1995 –2000 and 2004– 2007 periods . At the other end of the education spectrum, California saw a net increase in college -educated persons—both native -born and immigrant domestic migrants—during the 1985– 1990 and 1995 –2000 periods . Even with a net loss in college-educated native-born migrants during 2004 –2007, California still had a net increase in college -educated immigrants who moved within the United States from 2004 to 2007. Combining this in-migration of college - educated immigrants and out-migration of high school dropout immigrants produces an overall increase in education among immigrant domestic migrants to California . Newly arriving immigrants, who drive the trends for immigrants overall, are increasingly likely to have a college degree over these three time periods and decreasingly likely to have less than a high school diploma . This is the first time in recent California history that the proportion of new highly educated immigran ts exceeded that of less educated immigrants. California’s new immigrant arrivals appear to look different now in terms of education level . Changes in the educational composition of California’s immigrants may forecast changes in resource needs fo r the state, for example bilingual education . But education is only one characteristic of new immigrants. Understanding the new trends in immigrant settlement—and the implications for the state—requires a look at additional socioeconomic dimensions . 19 Changes in the Composition of California’s New Immigrant Arrivals As shown above, new immigrants to California have higher levels of education, on average, in 2007 than in 2000 or 1990. We next explore whether this trend is unique to Califor- nia , consider characteristics of new immigrants in addition to their education level s, describe the socioeconomic characteristics of new immigrants to California relative to new immigrants in other parts of the country , and examine how these characteristics have changed over time. Historically, immigrants have not been spread out evenly across the United States, but rather have t ended to cluster in a relatively small number of cities and states and often locate near other immigrants with similar socioeconomic character istics. For example, both Florida and California are popular immigrant destination states, but Florida is much more likely to attract Cuban immigrants than California is, and California is a more likely destination for Mexican immigrants than Cuban ones . O ver time, however, this clustering has declined, an d markedly so for California . We use statistical models that relate an immigrant’s individual characteristics to the choice of living in California relative to other states in each of the years 1990, 2000, and 2007. 24 These models reveal how immigrants’ socioeconomic characteristics, on average, are related to their choice of living in California rather than other states . For each year, we estimate the likelihood of new immigrants choosing to live in the state given a large set of characteristics, such as age, education, ethnicity 25, gender, marital status, English fluency, employment, wage level, industry of employment, housing costs, and homeownership. 26 The models also show that although new immigrants to California look very different from those moving to other states, in socioeconomic terms this disparity has declined since 1990. 27 24 These are simple linear probability models. Probit estimates are similar. Figure 3 shows the estimated probability that an immigrant chose to live in California in 1990, 2000, or 2007, and shows how that probability changes across immigrants’ characteristics. (The chart displays only a handful of characteristics included in the statistical model; for full results see Technical Appendix Table A4. ) A typical immigrant wa s about 37 percent more likely to choose California than other states in 1990 (“Average” column ). This probability decline d to about 18 percent in 2007. If an immigrant had less than a high school degree, he or she was even more likely to choose the state (comparing the column “Less than High School” to the horizontal line ). While less educated new immigrants a re more likely to live in California than other states, this clustering of less educated new immigrants in California has declined over time. An immigrant with less th an a high school degree wa s about 40 percent more likely to choose California in 1990 . By 2007, this wa s down to 23 percent. Conversely, college - educated new immigrants were less likely to choose California, but the difference decreased over time (“College Degree” column) . 25 Throughout this report we define ethnicity variables based on self -reported “race” categories given in the Census samples. They are as follows: Latino, non -Latino White, non-Latino Black, and Asian. All others are included in a final “other” category. 26 Technical Appendix B provides more detail on some of these variables. 27 See Technical Appendix Table A4 for detailed regression results. 20 Figure 3 . Predicted Probability of Living in California for New Immigrants with Selected Characteristics SOURCE: Author’s calculations from IPUMS Decennial Census and American Community Survey data. NOTES: Includes only imm igrants aged 18-64 with 0-5 years in the United States. Bars represent predicted probability from linear probability model described in text. “Average” is the predicted probability at the average value of explanatory continuous variables and taking account of the distribution across categorical va riables. All other bars use these average values except for the characteristic listed . By comparing the bar to the horizontal lines for 1990 or for 2000 and 2007, one can estimate the percentage point difference in predicted probability for each factor compared to the average predicted probability in that year . Another notable feature of the immigrant clustering in California is its ethnic dimension. A new Latino or Asian immigrant wa s more likely to live in California than other states in 1990, Latinos roughly 20 percent more likely , and Asians 1 4 percent more likely. But this overrepresentation has decline d over time. By 2007, a Latino new immigrant wa s only 5 percent more likely to live in California a nd an Asian new immigrant 13 percent. To see this graphically, we compare immigrants who are either Latino or Asian to all immigrants. The probability of choosing California was still high, but declined markedly over time, especially fo r Latino s. Compared to average, a Latino immigrant was about 7 percent more likely to choose California than other states in 1990. By 2007, a Latino immigrant was no more or less likely than an average immigrant to choose California . This trend for Latino immigrants in particular stands in contrast to the state’s history of attracting a large percentage of new Latino immigrants. New immigrants to California cluster not only by education and ethnicity, but also along some economic dimensions. In 1990, new immigrants to California were much more likely to be 0 5 10 15 20 25 30 35 40 45 50 Av er a geLa t in oAsianLess than High School College Degr ee Co n s t ruct ion Industry Ma n u fa cturing Industry Predicted Probability New Immigrant with Average Characteristics Chooses Calif ornia 199020002007 36 20 19.8 Probability for Avg Immigrant in 1990 Probability for Avg Immigrant in 2000 and 2007 21 employed in agriculture, construction, manufacturing, and in some service industries than were new immigrants to other states. Each of these industries, except manufacturing, employed a larger fractio n of the labor force in California than in other states (see Te chnical Appendix C). The predominance of these industries is likely a draw for new immigrants, but the industry composition of employment in the state is also affected by the employment of new immigrants. The likelihood of new immigrants employed in construction and manufacturing industries to choose California over other states in 1990 is shown in the height of the darkest bars in the Figure 3 Construction and Manufacturing columns. For new imm igrants in construction, by 2000 , they we re more likely to live outside California (shown in the fall of predicted probability in the Figure 3 Construction column). Note that this coincides with an overall drop in the share of Californians working in const ruction. In manufacturing, by 2007, new immigrants in California were about as likely to work in this industry as we re immigrants in other states. We see similar declines in new immigrants in California working in business services, despite an increase in the percent of the California economy employed in business services (Technical Appendix C) . Thus, changes in California’s industry composition cannot be the only draw for new immigrants to the state. In su mmary, in 1990, 2000, and 2007, we find significant clustering of new immigrants in California along social and economic dimensions . New immigrants with less than a high school degree or Latino new immigrants are more likely to live in California than other states, and those with a college degree are less likely . However, the degree of this clustering has declined over time. The changes in characteristics of the new immigrants to California along ethnic and education dimensions in particular stand in contrast to a long historical trend of immigration to the state. 22 Choices within California Although the large inflow of new immigrants to California continues, more new immigrants to the United States are choosing to live in other states. Similarly, it appears that as immigrant populations grow in new destination states, they are also growing in new areas of this state. Nationally, new immigrants who chose California were like new immigrants to other states along socioeconomic dimensions after 2000 tha n they were before 2000. In this section we examine whether the same can be said for trends among new immigrant s to different areas wit hin California . Within California, Los Angeles County receives the most new immigrants but also has the largest outflow of immigrants to other counties or states. Table 6 shows the components of population change in some of the major counties in the state. 28 F our counties always rank at the top based on immigrant population since 1990: Los Angeles, Orange, San Diego, and Santa Clara . In all of these historically immigrant -rich counties except Santa Clara, the number of new immigrants fell from the 1985 –1990 period to the 1995 –2000 period and rebounded in the 2004– 2007 period, on a per -year basis. 29 Most of California’s gateway counties experienced a net outmigration due to internal relocation since 1985. In all gateway counties except Santa Clara, net outmigration accelerated from 1985 to 2007. Over the latter two migration periods we measure, the outmigration is not just of native-born but also of immigrants. To some extent this mirrors the slowing of new immigrant arrivals to these areas. This finding runs counter to research on earlier periods arguing that primarily native -born persons respond to immigrant influxes by relocating. However, in all the gateway counties, the number of new arrivals was vastly larger than the increase from internal migration . This is consistent with the nationwide trend that in gateway areas, new immigrant arrivals drive the change in immigrant populat ion. 30 28 See Technical Appendix Table A6 for sta tistics on all counties in California. It suggests that the factors underlying the changes in settlement patterns for immigrants are common factors to the location decisions for native-born individuals . 29 For the comparison of 2004 –2007 to earlier periods, it is necessary to consider the population change on a per year basis; 1985– 1990 and 1995– 2000 changes are divided by 5, and 2004-2007 divided by 3. Given the available data, this is the best comparison that we can do. However, the shorter migration period of 2004 –2007 may induce some bias. 30 Frey and Liaw (1998), for example. 23 Table 6. Components of migration in Calif ornia counties 1985–1990 1995–2000 2004–2007 Net internal migration of: Net internal migration of: Net internal migration of: County Net internal migration = Native born + Immigrant New immigrant arrivals Net internal migration = Native born + Immigrant New immigrant arrivals Net internal migration = Native born + Immigrant New immigrant arrivals Top immigrant destinations Los Angeles -283,810 -225,756 -58,054 440,861 -306,646 -149,416 -157,230 319,314 -294,952 -159,689 -135,263 220,417 Orange -23,968 -32,454 8,486 92,385 -31,250 -19,955 -11,295 84,720 -57,614 -36,616 -20,998 63,990 San Diego 43,327 31,107 12,220 62,647 -33,456 -27,578 -5,878 58,845 -68,085 -60,240 -7,845 57,687 Santa Clara -41,565 -42,747 1,182 59,234 -51,025 -50,587 -438 95,280 -16,397 -20,973 4,576 67,659 Top growing immigrant destinations Alameda -19,763 -25,915 6,152 36,257 -4,975 -22,136 17,161 59,450 -13,265 -10,369 -2,896 42,720 San Bernardino 119,517 92,104 27,413 24,103 -2,358 -14,525 12,167 23,161 23,546 -513 24,059 21,831 Riverside 131,698 103,769 27,929 21,067 48,212 36,798 11,414 24,017 103,959 53,288 50,671 25,142 Sacramento 41,928 36,247 5,681 12,003 5,547 1,086 4,461 23,903 1,327 1,443 -116 23,294 Kern 11,636 10,315 1,321 8,089 -12,934 -10,037 -2,897 8,244 16,776 9,137 7,639 8,628 SOURCE: Author’s calculations from IPUMS Decennial Census and American Community Survey data. NOTES: Includes immigrants age d 23 –64 as of 1990, 2000, or 2007, so as to count only persons of working age over the entire migration period. 24 Alameda, San Bernardino , Riverside, Kern and Sacramento Counties are fastest growing in the state in terms of immigrant population. 31 The re are two distinct patterns of growth , shown in the bottom panel of Table 6 . 32 In Alameda and Sacramento counties, growth in immigration comes predominantly from growth in new immigrant arrivals. These areas might thus be deemed new immigrant gateways. In San Bernardino and Riverside C ounties, however, more of the growth comes from internal migration of older arrival immigrants. During the 1985– 1990 and 2004 –2007 periods , more immigrants moved into San Bernardino and Riverside from domestic locations than from other c ountries. These counties are adjacent to Los Angeles County, and th e previous inflow of immigrants was due primarily to immigrants relocating from Los Angeles. 33 Just as immigration increases to new destination states have spurred legislation , so have the increases in local areas within California with little history of immigration . For example, San Bernardino County ha s made national news with a proposal for restrictive ordinances to regu - late immigration. These findings indicate that Los Angeles continues to function as a gateway for immigrants who subsequently move elsewhere, especially to nearby inland counties . Kern C ount y’s migration pattern lies in between the two, with little growth in new immigrant arrivals, but a significant inflow of older arrival immigrants only from 2004–2007. 34 Figure 4 shows the education distribution of immigrants in six state MSAs . In 1990, 42 percent of the immigrant population of the Los Angeles MSA had not fi nished high school, but by 2000 that ratio was down to 39 percent and to only 32 percent by 2007. Many MSAs in the state experienced increases in average education but none at as high a rate as Los Angeles. Similarly, most MSAs in the state—even new growth areas like Sacramento —experienced an increase in the percent of immigrants who had a college degree. Not surprisingly , given the nature of their industries, the San Francisco and San Jose MSA s have the highest fraction of immigrants with college degrees, 35 percent and 45 percent in 2007, respectively , up from 26 percent and 29 percent in 1990. Such action could be sparked simply by overall growth of the immigrant population in areas with little history of incorporating immigrants, but could also be related to particular socioeconomic changes driven by new immigrant arrivals. 35 Unlike new -growth states across the country, many of which experienced an increase in the fraction of immigrants with less than a high school diploma, new-growth MSAs in Cali- fornia saw declines in the share of these immigrants, similar to the trend for the state overall. 31 As noted previously for high growth states, we define high immigrant growth counties as the top 10 in terms of growth over 1990 –2000 and at least in the top 20 in terms of number of immigrants as of 1990. 32 To view these migration statistics as rates per number of residents, so as to compare the changes across areas of different size, see T echnical Appendix Table A7. 33 Results available upon request. 34 Chang, Cindy, “California City Council Rejects Anti -Immigration Legislation,” New York Times, May 16, 2006. National Public Radio, Morning Edition, “California Town Aims to Bar Illegal Immigrants from Renting,” May 12, 2006, accessed 1/22/09 at: http://www.npr.org/templates/story/story.php?storyId=5400392. 35 Santa Clara County has a high proportion of its workforce employed in manufacturing relative to other counties and San Francisco ci ty-county has a high proportion employed in retail trade, finance, and professional services. See T echnical Appendix C. 25 Figure 4. California MSAs Immigrant Education Distribution SOURCE : Authors’ calculations from IPUMS Decennial Census and American Community Survey data. NOTE : Includes only immigrants age d 18– 64. As in the previous section, we examine changes in socioeconomic characteristics of immigrants across California in a multidimensional manner , rather than by examining education levels alone . Earlier , we used statistical models to describe the characteristics of new immigrants who choose California relative to those choosing other states. Since Los Angeles County accounts for about half of all immigrants in the state, and because the decline in immigration to Los Angeles drives the decline in the overall state immigrant population, we will examine the choice of Los Angeles over all other areas in the state. As above, we estimate 42% 24% 19% 39% 24% 32% 27% 18% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates Los Angeles-Long Beach, CA 23% 31% 20% 35% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates San Francisco-Oakland-Vallejo, CA199020002007 44% 24% 11% 43% 26% 19% 12% 37% 30% 19% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates Riverside-San Bernadino, CA 24% 20% 39% 17% 45% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates San Jose, CA 34% 24% 17% 31% 24% 28% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates San Diego, CA 28% 19% 27% 21% 26% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates Sacramento, CA 26 how new immigrants’ characteristics are related to the likelihood of choosing to live in Los Angeles compared to that of living elsewhere in the state. The models corroborate the findings on education distribution discussed above. 36 When we examined California overall, we found evidence that the industry of employ - ment for immigrants who choose the state has changed from 1990 to 2007 . In particular, the likelihood of immigrants who work in construction and manufacturing choosing California declined. However, when looking within the state at the choice of Los Angeles ove r other areas, we find the strongest clustering of new immigrants in Los Angeles in service industries. The draw of the entertainment and hospitality service industry in the Los Angeles MSA increases over the period of study . Similarly, the likelihood of n ew immigrants in the finance, insurance, and real estate industry in Los Angeles has increased sharply relative to other areas in the state. New immigrants who work in construction are less likely to live in Los Angeles in 2000 and 2007, and we also see de clines among immigrants who work in the manufacturing industry. The decline in new immigrants who work in manufacturing is consistent with the overall decline in the percent of the Los Angeles workforce in manufacturing from 1990 to 2007, but there was no similar decline in the overall percent of the Los Angeles workforce employed in service industries or construction. Latino new immigrants were 11 percent more likely to live in the Los Angeles MSA in 1990, but only 6 percent more likely in 2007 . New immigrants who decide to live in Los Angeles are more likely to lack a high school diploma , but that is less so over time. New immigrants in Los Angeles earn less in wages than other new immigrants in the stat e, after controlling for indiv- idual characteristics related to earnings potential , moving from about 1 percent less in 1990 to 6 percent less in 2000 and 2007. 37 Immigrants’ location decision s are not only dependent on their characteristics but also on the conditions of the different geographic areas the y have to choose from. In the next section, we combine the analysis of the characteristics of new immigrants to California and the charac - teristics of the locations they choose from in order to uncover some of the factors behind these immigrants’ settlement decisions. 36 See Technical Appendix Table A8 for regression results. As before, we execute a linear probability model of living in the Los Angeles MSA relative to elsewhere in the state. 37 See Technical Appendix C for details. 27 Have New Immigrants’ Preferences Changed? Immigrants choose specific destinations for a host of reasons . R esearch has shown the primary reason behind immigrant location choice is social ties, with economic factors a distan t second . However, in light of new settlement patterns and changes in the characteristics of immigrants choosing California, the importance of these factors may have changed . Immigrants to the United States and residents who consider relocating within the coun- try weigh many factors . There are a number of theories of migration and location choice, which give varying weight to determinants that fall roughly into the economic, social, and institutional categories. 38 Previous research finds that economic costs and benefits are also significant factors in immigrant location choice, but are second to social fact ors. Among measures used to analyze the economic draw of a locale are jo b opportunity, wage level, employment rate, and housing prices. Individual-level characteristics such as education, work skills, and occupational status also factor into location choice since they are strongly tied to economic opportunity . The relative importance of these economic factors is difficult to gauge. For example, the draw of high wages often coincides with high housing prices, which otherwise may be undesirable. Thus, a number of economic factors may work together in complex ways. The distance to one’s home country is related to the cost of a given location, not simply economic cost but also psychic costs. This factor is thus interpreted as both an economic and social factor and will be included in the statistical models here. Social factors for migration, in particula r the concentration of co-ethnics in a local- ity, have been found to be the strongest predictor of immigrant location choice. This measure is very broad, but is used as a proxy for social ties to a particular location, for job opportunities arising from social connections, and for other associated informational or social benefits . Some institutional factors that are postulated to affect migration include state welfare benefit generosity, border crossing locations, and broad changes in Unites States immigration policy . Analysis of these factors is limited by other economic and social changes that coincide; however , we are able to include a measure of state welfare benefit generosity . The model estimates the probability that a new immigrant to the United States chooses an MSA based on its social, economic, and institutional characteristics . To do this parsimoni - ously, I use a multinomial framework, where individuals choose among MSAs as a function of individual and MSA characteristics. 39 We focus on Latino and Asian immigrants only in these models; further detail on immi - grant ethnicity cannot be estimated due to limited sample size. The statistical models require a large sample of individuals in order to estimate the effects we are interested in accu rately. By To incorporate differences across immigrants, the models are estimated separately by ethnicity. To examine changes in the importance of the factors over time, the models are also estimated separately by year. Details on some of the more technical aspects of this location choice model and full re gression statistics are available in Technical Appendix D. 38 For further detail on the theory and findings of research on immigration location choice, see Technical Appendix A. 39 In particular, I use the conditional logit model of McFadden (1973). The probability P(i,j) that individual i chooses location j is a function of MSA characteristics Z and individual characteristics X. 28 looking at only these two gro ups, we cover roughly 70 percent of the new immigrant popula - tion. New Latino immigrants make up 43 percent of new immigrants from 2002 to 2007 and new Asian immigrants constitute 30 percent. We find that the trends in location choice vary between Latinos and Asian immigrants, so the results for each group are discussed separately . Have Latino Immigrant Preferences Changed ? First we examine the determinants of location choice among new Latino immigrants. I n Figure 5, w e see that in 1990, having co -ethnics in an MSA was a very strong draw for new immigrants. The percent of the MSA population of the same ethnicity is used to measure the presence o f co-ethnics in these models. A 10-point higher fraction of Latino immigrants in an MSA was related to a 25 - percent higher chance that new Latino immigrants would choose that MSA . Similarly, increasing distance from the home country wa s a deterrent. However, these factors have decline d in their importance over time. 40 We find little evidence that new Latino immigrants choose M SAs within states that have more generous welfare benefits. An additional $100 per year in TANF or AFDC benefits in an area was related to a 2.5 -percent higher chance that a new immigrant would choose t hat location in 1990, and to only a 1-percent chance in 2000 or 2007. By 2000, new Latino immigrants were not strongly influenced by proximity to others of the same ethnicity . In fact, a 10-point higher share of Latino immigrants in an MSA ma de it slightly less likely that new Latino immigrants would locate there. This marks a turnaround in the determinant which has historically predominated immigrants’ location decisions . Social factors are still very important in the location choice of new Latino immigrants —overall, a large percent of new Latino immigrants choose Los Angeles, for example —but in comparing two areas with different concentrations of co-ethnics, the relevance of this factor ha s fad ed. 41 40 To view the summary statistics on these explanatory variables, see Technical Appendix Table D2. The increase in average percent of Latinos across MSAs over time does not explain the fall in the coefficient estimates. Also, average distance to home country has declined, so if anything we would expect the coefficient on distance to home country to in crease over time. This relationship is difficult to identify, since state welfare generosity is related to other economic and social conditions that might attract immigrants. 41 An $100 per year increase is about a 1-percent increase at the average benefit level across states. See Technical Appendix Table D2. 29 Figure 5. Importance of Factors in Latino New Immigrant Location Choice SOURCE: Authors’ calculations from IPUMS Decennial Census and American Community Survey data. NOTE: Includes immigrants aged 23-64 as of 1990, 2000, or 2007, so as to count only persons of working age over the entire migration period, and immigrants with 0 -5 years residence in the United States. Bars represent estimated effect on probability of choosing an MSA given the changes in covariates listed on the x -axis. Only statistically significant results at the 1 -percent level are shown . Full model statistics are presented in Technical Appendix Table D1. Looking at economic determinants of location choice among new Latino immigrants, we find evidence that new immigrants are drawn to MSAs that have jobs available —the unemploy - ment rate was an important indicator over all periods . The higher the unemployment rate, the less likely a new Latino immigrant is to choose an MSA . For a 1 -percent higher unemployment rate in an MSA, a Latino immigrant in 1990 wa s 15 percent less likely to choose that location . By 2007, thi s was down to about 6.5 percent , al though still statistically significant . New Latino immigrants are drawn to MSAs with high average wages, but only in 1990. As noted, it is difficult to predict the effect of higher wages or higher housing prices on location decisions . We might expect that immigrants would choose an MSA with a low cost of housing, holding everything else constant. However, labor market conditions, especially wages, are related to housing market conditions and are a very strong draw . T he correlation of these variables makes it hard to separately identify the draw of low-cost housing , and this probably explains the relatively small effects. 42 42 Excluding housing or wages from the model has little or no effect on other estimated coefficients. -20 -15 -10 -5 0 5 10 15 20 25 30 10 pct pt higher concentratio n of co-e thnic s 100,000 more people$100 higher weekly wage1 pct pt higher une mp. ra te$100 mor e / mon t h max welfare benefit 100 mor e mile s to home c ou ntry$10,000 higher hou sing c os t inde x Increase or Decrease in Probability of Choosing MSA Change in MSA Characteristic of Interest 1985-1990 Latino Immigra nt Arriva ls in 1990 1995-2000 Latino Immigra nt Arriva ls in 2000 2002-2007 Latino Immigra nt Arriva ls in 2007 30 Have Asian Immigrants’ Preferences Changed ? In contrast to the results for new Latino immigrants, changes in the draw of social fac - tors for new Asian immigrants are mixed . Over time, n ew Asian immigrants are less likely to choose a destination due to proximity to their home countr ies. Because the average distance for an Asian immigrant to his or her home country is f arther than for a Latino immigrant, the decrease in probability due to distance to h ome country is much smaller. Asian immigrants are increasingly likely to choose an MSA because of the presence of other Asian immigrants there. Here agai n, we find no evidence that Asian immigrants choose their destinations be- cause of generosity of welfare programs. The change in probability of choosing an MSA because the area offers $100 more in welfare benefits is estimated at zero or nearly zero in all years . Economic factors are a similarly strong draw for new Asian immigrants. Asian immi- grants were 20 percent more likely to choose an MSA because it ha d $100 higher weekly wages, all else equal, in 2007 . This estimate for Asians is stronger than for Latino immigrants, where statistically significant effects of higher MSA wages were found only in 1990 . New Asian immigrants were strongly deterred from choosing an MSA with higher unemployment rates. 43 Last, the difficulty in estimating the effect of housing prices on immigrant location choice is evident here as well . It is unclea r from the estimates whether new Asian immigrants are consistently drawn to MSAs with higher housing prices or lower ones. We expect that this is likely driven by the relationship of local area wages and housing prices. The size of the estimates is consistent with the size estimated for Latino immigrants, although both show evidence that this factor has declined somewhat in importance over time. 43 Additional testing reveals that the deterrent effect of the unemployment rate increases as the rate increases. 31 Figure 6. Importance of Factors in Asian New Immigrant Location Choice SOURCE: Authors’ calculations from IPUMS Decennial Census and American Community Survey data. NOTE: Includes immigrants aged 23-64 as of 1990, 2000, or 2007, so as to count only persons of working age over the entire m igration period, and immigrants with 0 -5 years residence in the United States. Bars represent estimated effect on probability of choosing an MSA given the changes in covariates listed on the x -axis. Only statistically significant results at the 1 -percent level are shown . Full model statistics are presented in Technical Appendix Table D1. Taken together, this simple model leads to some striking conclusions. The first is the decline in social factors as a driver of location choice . We still observe a great amount of clustering of immigrants along ethnic and home country lines, but the evidence here suggests this factor may play less of a role over time. Indeed, it reflects the dispersion of immigrants to new destination areas. Second, we find no proof that w elfare programs are a prominent factor in the location choice of new immigrants. On the other hand, there is evidence that economic factors have played a strong and consistent role in the location decision s of new immigrants. For both new Latino and Asian immigrants, higher average wage has a positive effect on the probability of choosing an MSA and unemployment rate has a negative one. These changes may be interpreted in different ways with regard to the assimilation of new immigrants. New trends in immig rant location choices to new destinations that have little history of immigration present challenges to local areas. The social adaption of immigrants may be slowed because they are less likely to choose areas simply because of the availability of immigrant networks that help to facilitate assimilation . However, the long-term success and broad assimilation of immigrants is also related to economic progress. The econo mic oppor- -20 -10 0 10 20 30 40 10 pct pt higher concentration of co-ethnics 100,000 more people$100 higher weekly wage1 pct pt higher unemp. rate$100 more/mont h max welfare benefit 100 more miles to home country$10,000 higher housing cost index Increase or Decrease in Probability of Choosing MSA Change in MSA Characteristic of Interest 1985-1990 Asian I mmigrant Arrivals in 1990 1995-2000 Asian I mmigrant Arrivals in 2000 2002-2007 Asian I mmigrant Arrivals in 2007 32 tu nity that new destinations present may spur faster economic assimilation and long -term success of new immigrants. Economic motivations for location choice may mitigate the challenges to new destina - tion areas . To the extent that economic opportunity leads to economic success, new immigrants motivated increasingly by economic opportunity are more likely to contribute to, rather than tax the system. In California, the recent decline in the popularity among new immigrants of the state itself and of the Los Angeles region in particular may be understood better in light of the preferences of new immigrants. Proximity to Mexico and to large Latino and Asian immigrant populations have historically driven continued flow of new immigrants to the state. The decline in importance of these factors explains a good portion of the decline in the state’s popularity. But we also find that economic opportunities are consistently a strong factor behind immigrant location choice . Additional r esults suggest that the location decisions of highly educated Latino and Asian new immigrants are even more strongly related to economic conditions in MSAs. 44 44 See Technical Appendix Figures D1 and D 2. Thus , the state’s ability to attract a highly skilled immigrants for its workforce is linked to economic conditions in other places. 33 Figure 7. Trends in Unemployment, Wages, and Housing Prices in California and the United States With the decline of social factors as a draw for immigrants, the state’s economic condition has increasing importance for drawing immigrants, and drawing highly educated immigrants in particular. The current economic crisis will have an impact on the numbe r and characteristics of immigrants who choose to come to the United States and who choose to live in California . The recent upturn in unemployment rates is seen in the third panel of Figure 7 . The California and Los Angeles unemployment rates have increased more quickly in the last year . To the extent that the economic decline is more pronounced in California than in other areas, $0 $50,000 $100,000 $150,000 $200,000 $250,000 $300,000 $350,000 $400,000 $450,000 $500,000 198519871989199119931995199719992001200320052007 Median House Price Median MSA Home Prices California MSAs Los Angeles-Long Beach MSA All US MSAs SOURCES: OFHEO Housing Price Index, US Census Bureau. NOTES: Median house prices respresent average across MSAs; 1999 dollars. $0 $5,000 $10,000 $15,000 $20,000 $25,000 $30,000 $35,000 $40,000 $45,000 $50,000 198519871989199119931995199719992001200320052007 Average Annual Wage Annual Wages for All Workers and New Immigrants All California WorkersAll US WorkersNew California ImmigrantsNew US Immigrants SOURCE: Current Population Survey (US Census, Bureau of Labor Statistics). NOTES: 18 -64 year-oldm ale s o nly; immigration status not available prior to 1994; 1999 dollars. 0% 2% 4% 6% 8% 10% 12% 198519871989199119931995199719992001200320052007 Average Annual Unemployment Rate Average Annual Unemployment Rates California Los Angeles-Long Beach MSA US SOURCE:Bureau of Labor Statistics. NOTE: Average annual unemployment rates are not seasonally adjusted. 34 the state will likely be disadvantaged in attracting immigrants in the future. However, the first two panels of Figure 7 , showing median housing prices and wage rates, are not definitive in predicting how California’s draw for new immigrants might change. Even with the economic downturn, California has a higher average wage than the rest of the United States, a draw , and the state’s markedly higher housing prices , a deterrent, have started to decrease. 35 Conclusion Although California is still home to more immigrants than any other state, its popularity for immigrants began to wane for the first time after 1990 . Calif ornia ’s 44-percent immigrant growth rate from 1990 to 2000 was far exceeded by growth rates of more than 200 percent in states such as Georgia and North Carolina . Within California, the number of immigrants in Los Angeles County increased an average of ju st 1.9 percent per year during the 1990 –2007 period , compared to rates as high as 12.6 percent per year in Riverside County and 10.5 percent in Kern County. In this report, we find that most of the change is driven by the changing location choice of new arrival immigrants rather than those of previous immigrants. Within California there is also a significant shift of immigrants to new destinations within the state, although the traditional immigrant gateway cities of Los Angeles and San Francisco still serve as jumping off points for new immigrants. Thus, programs to facilitate new immigrant assimilation in gateway areas are likely to continue to have a large population to serve. The trend towards settlement in new destinations is also driven in many areas by the internal migration of native -born persons, suggesting that determinants of location choice are increasingly common to both groups . Some researchers have shown that in the recent past, native -born persons tended to leave areas with high in -migration of immigrants. We find a reversal of this trend . In addition , our findings suggest that immigrant inflows are not responsible for initiating native -born outflows. Thus , it may be that the level of competition between native-born and immigrants has been o verstated. New Latino immigrant arrivals to the United States were much less likely to choose to live in California in 2007 than they were in 1990; new immigrants employed in construction, manufacturing, and some service industries also prefer other states. New immigrants choosing California over other states were slightly more educated in 2007 than in 1990 . We find that changes in the composition of new immigrants are related to the factors by which they make their decisions on wh ere to settle. Social factors , historical ly the principal determinant , have waned in importance for Latino immigrants in particular . New Latino immigrants are much less likely to choose cities because of the presence of co -ethnics. These social factors are still the primary explanation for location choice, but economic factors are a strong second . This explains the decision of many immigrants to live in new destinations that h ave a less established immigrant social network but have growing economic opportunities. The established immigrant networks in California have less attraction for the newest immigrants to the United States, and this explains most of the decline in the state’s popularity. The waning importance of social factors in new immigrants’ locati on decisions may signal the assimilation of immigrants . For those concerned with the integration of immigrants, the decline in clustering of immigrants along social lines and the increasing i mportance of economic factors may be good news . On the other hand, in areas with few immigrants and little experience incorporating immigrants into social and economic life, the settlement of immigrants away from co -ethnics raises new challenges. 36 References Bartel, Anne , “Where Do the New U.S . Immigrants Live?” Journal of Labor Economics, V ol . 7, No. 4 , 1989. Bohn, Sarah, “Immigration and Wages in the U.S. Labor Market ,” University of Maryland Dissertation, UMI Number 3283414 , 2007. Borjas, George , “The Labor Demand Curve Is Downward Sloping : Reexamining the Impact of Immigration on the Labor Market, ” Quarterly Journal of Economics, V ol . 118, No. 4 , 2003. Borjas, George , “Immigration and Welfare Magnets,” Journal of Labor Economics, Vol. 17, No. 4 , 1999. Borjas, George , and Bernt Bratsberg , “Who Leaves? The Outmigration of the Foreign -Born,” The Review of Economics and Statistics, V ol . 78, No. 1 , 1996 . Frey, William , and Kao -Lee Liaw , “The Impact of Recent Immigration on Population Redistribution Within the United States,” in James P. Smith a nd Barry Edmonston, eds., The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration, National Academy Press, Washington D .C ., 1998. Gibson, Campbell, and Emily Lennon, “Historical Census Statistics on the Foreign -Born Population of the United States 1850 -1990, ”U.S. Census Bureau Population Division Working Paper No. 29, 1999. Hopkins, Daniel, “Threatening Changes: Explaining Where and When I mmigrants Provoke Local Opposition,” Center for the Study of American Politics, Yale University , 2007. Jaeger, David , “Local Labor Markets, Admission Categories, and Immigrant Location Choice,” IZA Discussion Paper, 2000. Jasso, Guillermina , and Mark Rosenzweig , “Estimating the Emigrati on Rates of Legal Immigrants Using Administrative Survey Data: The 1971 Cohort of Immigrants to the United States, ” Demography , Vol. 19, No. 279–290, 1982. Kaushal, Neeraj , “New Immigrants’ Location Choices: Magnets without Welfare,” Journal of Labor Economics, Vol. 2 3, No. 1, 2005 . Lewis, Ethan, “Immigration, Skill Mix, and the Choice of Technique,” Federal Reserve Bank of Philadelphia Working Paper No. 05- 8, 2005. McFadden, D aniel, “Conditional Logit Analyses of Qualitative Choice Behavior ,” in P. Larembka, ed., Frontiers of Econometrics, Academic Press, New York, 1973. Massey, Douglas, Jorge Durand, and Nolan Malone , Beyond Smoke and Mirrors: Mexican Immigration in an Era of Economic Integration, Russell Sage Foundation, New York, 2002. Massey, Douglas, ed. , New Faces in New Places: The C hanging Geography of American Immigration, Russell Sage Foundation , New York, 2008 . 37 National Conference of State Legislatures, “2007 Enacted State Legislation Related to Immigrants and Immigration”, available at www.ncsl.org/programs/immig/2007immigrationfinal.htm, Jan. 31 2008 . National Conference of State Legislatures, “State Laws Related to Immigrants and Immigration: January 1 – June 30, 2 008”, available at www.ncsl.org/programs/immig/immigreportapril2008.htm, July 28 , 2008. National Conference of State Legislatures, “State Laws Related to Immigrants and Immigration: January 1 – June 30, 2008”, available at www.ncsl.org/programs/immig/immigreportjuly2008.htm, July 24 , 2008. National Conference of State Legislatures, “State Laws Related to Immigran ts and Immigration in 2008”, available at www.ncsl.org/programs/immig/2008StateLegislationImmigration.htm, January 27 2009. Passel, Jeffrey and Wendy Zimmerman , “Are Immig rants Leaving California? Settlement Patterns of Immigrants in the Late 1990s, ” Urban Institute, 2001. Peri, Giovanni , “How Immigrants Affect California Wages and Employment, ” Public Policy Institute of California , San Francisco , 2007. Ramakrishnan, S. K arthick, and Paul Lewis , “Immigrants and Local Governance: The View from City Hall ,” Public Policy Institute of California , San Francisco , 2005. Reed, Deborah , “California’s Future Workforce: Will There Be Enough College Graduates ?” Public Policy Institute of California , San Francisco , 2008. Ruggles, Steve, Matthew Sobek, Trent Alexander, Catherine A. Fitch, Ronald Goeken, Patricia Kelly Hall, Miriam King, and Chad Ronnander , Integrated Public Use Microdata Series: Version 3.0 [Machine -readable database] . Minneapolis : Minnesota Population Center [producer and distributor] , 2004. Singer, Audrey , “The Rise of New Immigrant Gateways,” The Brookings Institution Center on Urban and Metropolitan Policy, Washington, D.C., 2004. Singer, Audrey, Susan Hardwick, and Caroline Brettell, eds. , Twenty -First Century Gateways: Immigrant Incorporation in Suburban America, Brookings Institution Press, Washington, D.C., 2008 . Singer, Audrey , and Robert o Suro , “Latino Growth in Metropolitan America: Changing Patterns, New Loc ations,” The Brookings Institution , Metropolitan Policy Program, 2002. U.S. Department of Homeland Security, Task Force on New Americans, “ Building an Americanization Movement for the Twenty -first Century: A Report to the President of the United States fro m the Task Force on New Americans, ” Washington, D.C., 2008. Zavodny, Madeline , “Welfare and the Locational Choices of New Immigrants,” Federal Reserve Bank of Dallas Economic Review, Second Quarter, 1997. Zavodny, Madeline , “Determinants of Recent Immigrants’ Locational Choices,” International Migration Review, Vol. 33, No. 4, 1999. 38 About the Author Sarah E. Bohn is a research fellow at the Public Policy Institute of California, where she studies immigrants, immigration policy, economic demography, and the economy. Her work focuses specifically on the effects of immigration on the labor market and on the economic progress of immigrants in the United States. She holds a Ph.D. in economics from the University of Maryland, College Park. PUBLIC POLICY INSTITUTE OF CALIFORNIA Board of Directors Walter B. Hewlett, Chair Director Center for Computer Assisted Research in the Humanities Mark Baldassare President and Chief Executive Officer Public Policy Institute of California Ruben Barrales President and Chief Executive Officer San Diego Regional Chamber of Commerce John E. Bryson Retired Chairman and CEO Edison International Gary K. Hart Former State Senator and Secretary of Education State of California Donna Lucas Chief Executive Officer Lucas Public Affairs Ki Suh Park Design and Managing Partner Gruen Associates Constance L. Rice Co -Director The Advancement Project Thomas C. Sutton Retired Chairman and Chief Executive Officer Pacific Life Insurance Company Raymond L. Watson Vice Chairman of the Board Emeritus The Irvine Company Carol Whiteside President Emeritus Great Valley Center PUBLIC POLICY INSTITUTE OF CALIFORNIA 500 Washington Street, Suite 600 San Francisco, California 94111 p hone: 415. 291.4400 f ax: 415. 291.4401 PPIC SACRAMENTO CENTER Senator Office Building 1121 L Street, Suite 801 Sacramento, California 95814 p hone: 916.440.1120 f ax: 916.440.1121 www.ppic.org" } ["___content":protected]=> string(102) "

R 709SBR

" ["_permalink":protected]=> string(93) "https://www.ppic.org/publication/new-patterns-of-immigrant-settlement-in-california/r_709sbr/" ["_next":protected]=> array(0) { } ["_prev":protected]=> array(0) { } ["_css_class":protected]=> NULL ["id"]=> int(8647) ["ID"]=> int(8647) ["post_author"]=> string(1) "1" ["post_content"]=> string(0) "" ["post_date"]=> string(19) "2017-05-20 02:39:29" ["post_excerpt"]=> string(0) "" ["post_parent"]=> int(3911) ["post_status"]=> string(7) "inherit" ["post_title"]=> string(8) "R 709SBR" ["post_type"]=> string(10) "attachment" ["slug"]=> string(8) "r_709sbr" ["__type":protected]=> NULL ["_wp_attached_file"]=> string(12) "R_709SBR.pdf" ["wpmf_size"]=> string(6) "537102" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(89212) "New Patterns of Immigrant Settlement in California Sarah Bohn with research support from Eric Schiff Supported with funding from The Ford Foundation and from t he Research Foundation of The City U niversity of New York July 2009 The Public Policy Institute of California is dedicated to informing and improving public policy in California through independent, objective, nonpartisan research on major economic, social, and political issues. The institute’s goal is to raise public awareness and to give elected representatives and other decisionmakers a more informed basis for developing policies and programs. The institute’s research focuses on the underlying forces shaping California's future, cutting across a wide range of public policy concerns, including economic development, education, environment and resources, governance, population, public finance, and social and health policy. PPIC is a private operating foundation. It does not take or support positions on any ballot measures or on any local, state, or federal legislation, nor does it endorse, support, or oppose any political parties or candidates for public office. PPIC was established in 1994 with an endowment from William R. Hewlett. Mark Baldassare is President and Chief Executive Officer of PPIC. Walter B. Hewlett is Chair of the Board of Directors. Copyright © 2009 by Public Policy Institute of California All rights reserved San Francisco, CA Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source and the above copyright notice is included. Research publications reflect the views of the authors and do not necessarily reflect the views of the staff, officers, or Board of Directors of the Public Policy Institute of California. 3 Contents Summary 4 Acknowledgments 5 Introduction 6 CALIFORNIA’S POPULARITY DECLINE 13 CHANGES IN THE COMPOSITION OF CALIFORNIA ’S NEW IMMIGRANT ARRIVALS 19 CHOICES WITHIN CALIFORNIA 22 HAVE NEW IMMIGRANTS’ PREFERENCES CHANGED? 27 Have Latino Immigrant Preferences Changed? 28 Have Asian Immigrants’ Preferences Changed? 30 Conclusion 35 References 36 About the Author 38 4 Summary California has been a crucible for immigrant -related issues in the United States for decades due to the overwhelming share of U.S. immigrants who choose to live in the state . However, in the late 1990s, the popularity of California among immigrants began to decline for the first time in nearly 100 years. As fewer immigrants locate to California and to other traditionally immigrant -rich areas, more are choosing to live in states and cities with little history of immigration. One consequence of this demographic change is a rise in immigration -focused local legislation in those new settlement areas. In the first half of 2008 alone, more than 1,200 state bills related to immigration were proposed across the country . Immigration is no longer an issue limited to a n isolated handful of states or a handful of cities; it affects many areas across the country. In light of this demographic shift and associated policy responses, and to provide context for the immigration debates being waged nationwide and locally, this study examines immigration in detail using U.S. Census data from 1990 –2007 . We find that the decline in California’s share of the nation’s immigrant population is driven partly by out -migration of established immigrants to other states, but mostly by the settlement of new immigrant arrivals in to different states. California has experienced a net out- migration of both established immigrants and native-born persons to other states, but the flow of established immigrants is relatively small. In contrast, California’s share of new immigrant arrivals to the United States has fallen sharply, from 35 percent of new arrivals in the late 1 980s to only 19 percent in 2004 –2007. Although new immigrants to the United States and to California are not markedly different from their predecessors, it appears that some of the socioeconomic considerations related to their migration choices have changed . Social factors such as residence near co -ethnics explain a large portion of immigrant concentration in California, but these have waned in importance . At the same time, economic factors have remained consistent in explaining where immigrants tend to locate. For those concerned with the integration of immigrants, the decline in clustering of immigrants along social dimensions may be good news . The relevance of economic opportunity in immigrants’ location decisions ma y also bode well for their economic integration . On the other hand, in areas with few immigrants and little experience incorporating immigrants into social and economic life, the settlement of immigrants away from co -ethnics may lead to increasing isolation for these groups and new challenges for the communities they settle in . All technical appendices to this paper ar e available on the PPIC website: http://www.ppic.org/content/pubs/other/709SBR_appendix.pdf 5 Acknowledgments Eric Schiff provided excellent research support for this project. The author acknowledges the helpful reviews of Hans Johnson, Laura Hill, Audrey Singer, and Randy Capps, as well as input from Dylan Conger, Hector Cordero -Guzman, Jason Fitchener , seminar participants at the Association for Public Policy Analysis and Management 2008 Fall Conference , and at the American Economic Association 2009 Annual Meeting. 6 Introduction By 2007, an estimated 38 million immigrants resided in the United States. About 26 pe rcent of those lived in California . For 90 years, the popularity of California as a destination for immigrants steadily increased . But in the 1990s, for the first time since the early 1900s, Cali - fornia’s draw for immigrants began to wane. The rapid immigrant growth over most of the state’s history has helped shape its demo - graphic and economic makeup, and has fueled fierce debate, as epitomized by Proposition 187 in 1994 and Proposition 227 in 1998. 1 Figure 1 . Percent of U.S. Immigrants Living in California, 1850 –2007 The marked turnaround in growth in the 1990s raises a host of questions. Among them are: why did the decline occur and what does it mean for California’s future? SOURCE: Passel and Zimmerman (2001), US Census Bureau (1999), author's calculations from Census data. California has a unique history with immigration, in terms of the number of immigrants who choose to live in the state and because of its consequent policy experience with immigrant- related issues. However, the decline in popularity of the state as an immigrant destination is not unique ; other leading immigrant destination states have experienced this decline as well . As Map 1 shows, the states with the largest concentration of immigrants in 2000 ex perienced some of the smallest increases in the number of immigrants from 2000–2007. 2 1 Proposition 187 was a broad -ranging measure dealing with immigrants in California that empowered all law enforcement agents to enforce federal immigration law and imposed restrictions on public benefits for immigrants. Proposition 227 eliminated bilingual education. California, New York, Illinois, Texas, Florida , and New Jersey have the largest immigrant 2 This map counts only immigrants of working age, 18 –64. See Singer (2004) for a similar map with all immigrants included for 1990-2000. 1.03.53.84.44.03.64.35.47.68.0 10.2 13.8 18.3 25.4 32.7 28.5 26.4 0 5 10 15 20 25 30 35 40 1850186018701880189019001910192019301940195019601970198019902000 Percent Ye a r P e rc e nt in CA (va lue s shown) P e rc e nt in CA of working a ge P e rc e nt of US P opula tion in CA 7 populations, but higher growth is found in n ew immigrant destination states. The same is true for immigration during the 1990– 2000 period. Map 1. Percent Change in Number of Immigrants , 2000 –2007 SOURCE: Authors’ calculations from IPUMS Decennial Census data. NOTE: Includes only immigrants age d 18-64 . Shading represents quintiles of the percentage change. The dispersion of immigrants across the United States is a relatively new phenomenon, and has been noted by many researchers. In 1990, 74 percent of all immigrants of working age in the United States lived in six states. Californi a alone was home to nearly 33 percent of them. This concentration had increased consistently over the previous 100 years of immigration history in the United States. 3 Throughout this report, an immigrant is def ined as a person born outside the United States and its territories and either a naturalized American or non -citizen. The documentation status of immigrants is a particularly heated aspect of the debate about immigration nationally and locally . T here are n o comprehensive data that allow researchers to identify the documenta - tion status of immigrants at the individual level, so this study will consider all immigrants For the first time, in the late 1990s, the percent of immigrants living in the top six states declined , falling to 69 percent in 2000 and to 66 percent in 2007. Although these decreases seem small against the backdrop of nearly 100 years of increasing concentration , the trend reversal is striking. California continues to be home to the largest number of immi -grants and has the highest concentration of immigrants (the ratio of immigrants to total popula -tion), but California rank ed 40th in the nation in percentage change in the number of immi -grants from 1990 to 2000 and 43rd from 2000 to 2007 . Between 1980 and 1990, California’s working -age immigrant population grew 9 .5 percent per year . This growth was down to 4 .4 percent per year between 19 90 and 2000 and to 2 percent per year between 2000 and 2007. 3 Passel and Zimmerman (2001). 8 regardless of legal status. I t has been argued that state and local policies aimed at illega l immigrants have spillover effects on all immigrants. 4 This demographic change in the United States overall raises important questions about how we understand the movement of population and how that movement contributes to the socioeconomic characterist ics of different areas. The academic literature on migration is well developed, but it has yet to fully analyze this re cent pattern. For California in particular, the decline in popularity of the state as an immigrant destination is not well understood. Have newly arrived immigrants simply decided to live in new places? Have established immigrants migrated away from California? And if so, are the same factors driving immigrants to new places affecting the location decisions of native -born California residents? These questions can be asked not only of immigrant settlement patterns between Cali - fornia and other states, but also of patterns within California . Within the state, Los Angeles County dominate s in number and concentration of immigrants. With nearly 2.3 million immi- grants in 1990 and 2.9 million in 2007, Los Angeles County has more than triple the number of immigrants of any other county . However, the county experienced very little growth in immi - grant population over this same period, especially compared to other counties . Between 1990 and 2006, the number of immigrants in Los Angeles County increased an average of just 1.9 percent per year, compared to rates as high as 12.6 percent annually in Riverside County and 10.5 percent in Kern County. Map 2 show s that the California counties with the largest immigrant populations experienced relatively low growth in immigrant population between 1990– 2000 and 2000 –2007. 5 Map 2 . Percent Change in Number of Immigrants, 2000–2007 SOURCE: Authors’ calculations from IPUMS Decennial Census data. NOTE: Includes only immigrants age d 18–64. 4 Singer et al (2008). 5 Census data allow us to identify 42 county groups in California. Small counties such as Sierra, Plumas and Nevada are grouped together because there is not enough information to accurately break out these counties individually. 9 Looking specifically at metropolitan areas within the state, we see that the Los Angeles Metropolitan Statistical Area (MSA) 6 Table 1 . Changes in immigrant population in California MSAs, 1990–2007 dominates all other state MSAs in terms of immigrant population . The Los Angeles MSA was home to 2.7 million immigrants of working age in 1990 and 3.7 million in 2007; this is about three times the size of the immigrant population in any other MSA . The San Francisco -Oakland MSA is the second la rgest, with about 629,000 in 1990 and 1.1 million in 2007. Number of immigrants Percent of MSA population Percent change MSA name 1990 2000 2007 1990 2000 2007 1990 –2000 2000 –2007 Bakersfield 48,246 84,365 129,227 15.7 23.7 28.7 74.9 53.2 Chico 8,031 10,998 13,774 7.6 9.3 10.0 36.9 25.2 Fresno 84,787 151,247 191,570 22.1 29.1 31.3 78.4 26.7 Los Angeles-Long Beach 2,693,286 3,493,571 3,654,601 37.7 46.0 45.6 29.7 4.6 Merced 25,458 38,174 49,491 25.8 33.0 34.7 49.9 29.6 Modesto 39,984 65,045 80,409 18.6 25.0 25.8 62.7 23.6 Redding 2,530 4,629 4,694 3.0 5.0 4.3 83.0 1.4 Riverside-San Bernardino 271,754 490,946 761,629 18.3 26.8 31.3 80.7 55.1 Sacramento 101,296 170,761 245,361 11.1 17.4 21.0 68.6 43.7 Salinas-Seaside-Monterey 57,542 62,833 64,319 28.3 40.0 39.1 9.2 2.4 San Diego 325,728 473,439 521,239 21.3 27.9 28.8 45.3 10.1 San Francisco -Oakland-Vallejo 628,890 967,173 1,078,199 23.9 32.6 35.4 53.8 11.5 San Jose 273,524 470,193 530,580 27.8 43.1 47.7 71.9 12.8 Santa Barbara-Santa Maria-Lompoc 47,123 66,564 68,661 20.6 27.9 29.0 41.3 3.2 Santa Cruz 25,257 39,623 37,710 17.6 24.1 23.2 56.9 -4.8 Santa Rosa-Petaluma 25,530 48,578 61,592 10.8 17.3 20.9 90.3 26.8 Stockton 54,521 86,697 130,597 19.8 27.3 32.8 59.0 50.6 Ventura-Oxnard-Simi Valley 87,987 125,388 150,836 21.3 27.6 30.6 42.5 20.3 Visalia- Tulare-Porterville 40,252 64,482 82,602 23.6 31.4 33.9 60.2 28.1 Yuba City 10,825 18,125 21,126 15.3 22.9 21.4 67.4 16.6 California MSAs 4,852,551 6,932,831 7,878,217 27.5% 35.5% 36.8% 42.9% 13.6% SOUR CE: Author ’s calculations from IPUMS Decennial Census and American Community Survey data. N OTE : Includes only immigrants age d 18- 64. San Diego, Riverside -San Bernardino, and San Jose MSAs are the next largest, but have only roughly half as many immigrants as the San Francisco-Oakland MSA. D espite the dominance of a few MSAs in terms of immigrant population, the MSAs that experienced the largest changes in immigrant population over the period were not those with a history o f attracting many immigrants (Table 1) . 6 There are 23 Metropolitan Statistical Areas (MSA) defined for California, which identify roughly the various centers of population and economic activity in the state. An MSA is defined around a population center and may be comp rised of single or multiple counties. Although they do not cover all of the land of the state, the population in MSAs in California comprises at least 95% of the total state population over 1990 -2006. For this reason, the trends for California’s MSAs mimic the trends described above for California counties. 10 The changing settlement patterns of immigrants have occurred at the same time as some immigrant -related developments in policymaking at the state and local level. In 2007, 1,562 immigration -related bills were introduced by state policymakers across the United States, about three times the number introduced in 2006. 240 of the bills introduced in 2007 were enacted . In the first half of 2008 alone, 1,267 bills were introduced and at least 175 signed into law. 7 Many state and local governments have taken policy action in response to the down - loading of immigran t-related issues and costs from the federal to the state and local level. This new legislation has covered a broad range of issues, including employment eligibility , human tr afficking, public benefits , and driving licens es. There is a lot of variation in the nature of these laws and ordinances: some are extremely restrictive while others are more accommodating . For example, Oklahoma’s HB 1804 makes it a felony to harbor or sh elter illegal immigrants and requires state and local law enforcement to enforce federal immigration laws. In contrast, California’s AB 976 prohibits landlords from asking about, or taking any action based on, a tenant’s immigration status. At the local le vel, it is likely that thousands of ordinances were proposed , although we are not aware of a source that has collected comprehensive information on them. 8 The federal government controls the number of immigrants legally allowed to enter the country each year, and to some extent controls the number of illegal immigrants th rough border control measures, arrests , and deportation s. State and local governments, however, have no direct power to regulate the number of immigrants who choose to settle within their borders. Thus, both the benefits and the costs of changes in immigrant population s accrue to local areas, but are largely out of the control of local governments . In the absence of federal immigration reform to assist local areas that receive large immigrant inflows, they have been left to deal with many of the challenges and costs on their own . This has lead to frustration at the local level, and a few lawsuits against the federal government. The policy levers that state and local governments can exercise range from those that deflect immigrants, such as enforcing zoning, licensing, an d housing codes, 9 and those that accommodate immigrants, such as establishing day labor sites and expanding bilingual education programs . This wide variety of responses is seen clearly in the range of state laws enacted in 2007 and 2008. 10 Previous studies have shown that changes in immigrant population, rather than its size, drive tensions at the local level. 11 7 Statistics in this paragraph are given in National Conference of State Legislatures January 31 and July 28 2008 reports on state legislation. The NCSL uses a comprehensive methodology for identifying all state legislation related to immigration. We are unaware of such a comprehensive methodology for local ordinances on the same subject. There are a variety of reasons. In non-gateway cities and states, the arrival of immigrants is a relatively new phenomenon . Since immigrants have hi storically settled in a very few number of places, there are a vast number of non -traditional immigrant destinations across the country that are ripe for tension. Indeed, the flurry of local legislation attempting to regulate immigration over the last few years reflects the potential for tension and conflict related to immigrant-driven demographic change. 8 Singer et al (2008), p. 157. 9 Light (2007). 10 While the effect and effectiveness of these policies with regard to immigrants is of great interest, not enough time has passed in order to fully identify the effects. We are planning future studies to carry out an evaluation of these policy changes. 11 Hopkins (2007), Singer et al (2008). 11 There is some evidence that areas with a long history of immigration become better at incorporating new immigrants. They may have established government services or community organizations offering assistance to immigrants. 12 In addition, ethnic enclaves have developed in many gateway cities that may both draw future immigrants and help them to become integrated, in particular by providing job opportunitie s. It may also be that over the long run, industries adapt to changes in population : the arrival of a large number of low skilled workers, for example, may induce a manufacturing firm to hire from the large pool of workers instead of investing in relatively more expensive machines that would otherwise replace workers. 13 Much of state regulation of im migration deals specifically with issues related to undocumented immigrants, including regulation of public benefits to illegal immigrants and restrictions on employment or housing . In California in 2007 –2008, 17 immigration -related laws were enacted and 1 5 resolutions were passed; nearly all of these laws and resolutions could be categorized as supportive of immigrants and immigration. Lastly, it may be that as fears about socioeconomic changes caused by the arrival of new immigrants are not realized, residents become less concerned . 14 The legislative surge relating to immigration issues begs the question of whether these new immigration patterns are related to the socioeconomic characteristics of immigrants or to the social and economic conditions in areas across the country. Immigrants have historically been clustered geographically, but they have also clustered around some socioeconomic characteristics . For example, immigrants in California are more likely to be Mexican, whereas immigrants in Florida are more likely to be Cuban. Similarly, t he San Francisco Bay a rea has tended to attract more highly educated immigrants, whereas the labor demand in other areas of the state attracts less educated immigrant workers. These dimensions of immigrant clustering affect local policy toward immigrants —for example, whether bilingual education programs or day labor centers are expanded —and they affect the demand for government-provided services and infrastructure . Thus, n ot only the fact that we experience new settlement trends is of interest, but also who is leaving, arriving, and staying becomes a concern . For example, if highly skilled new immigrants to the United States find California —or specific cities in California — less attractive, this trend may exacerbate the state’s problem of recruiting enough skilled workers. The same could be said of the six bills enacted in New York . The more accommodating nature of these bills may be evidence of these states’ long history of immigration . States with new immigrant growth tended to have a mix of legislative response to immigration issues. 15 Immigrant destination choice is also related to the success of immigrants themselves. We commonly frame the success of immigrants by their social and economic adaption to their destination of choice, commonly called assimilation. Historically, assimilation policies were often called “ Americanization” and largely referred to language learning, civic participation, naturalization, and acceptance of American cultural values (however defined) . B ut full immigrant assimilation also includes improving their economic outcomes with increased time Also, understanding the selectivity of migration may help policymakers estimate future funding needs for programs, services , and infrastructure. 12 Ramakrishnan and Lewis (2005). 13 Lewis (2005). 14 For additional detail on these bills, see Technical Appendix A. 15 Reed (2008). 12 spent in the United States and incorporation into social networks not related to becoming Americanized. Federal, state, and local policies aim to affect immigrant assimilation, for example, making bilingual education programs, job training, and migrant worker assistance services available . Local offices of i mmigrant affairs have long provided 16 We first examine the trend of immigrants locating to states other than California. Next, we look within the state at trends in the historically immigrant-rich are as and in new growth areas . Last , we provide an analysis of the underlying factors behind the change in immigrant location choice s and a discussion of the consequences. services to promote civic learning and to guide immigrants through the naturalization process. Recently, immigrant assimilation of this sort was a goal of former President George W. Bush’s Task Force on New Americans . Because the assimilation of immigrants is tied to their destination area and the reasons they choose to reside there, we explore the reasons behind immigrants’ changing settlement patterns. This may indicate prospects for immigrant assimilation and policies that might enhance assimilation. 16 California Senate Bill 1094 called for establishment of an Office of Immigrant Affairs. Other states and localities have similar offices with similar names. 13 California’s Popularity Decline Although California’s popularity as a destination among immigrants has declined since the late 1990s, the state is still home to the largest immigrant population in the country and that population has continued to grow . But t his growth in immigrant population in the state is much smaller than in the past and is much smaller compared to most states in the country . Table 2 shows the changes in immigrant population across states and groups of states durin g the period 1990 –2007. Table 2 . Changes in immigrant population, 1990– 2007 Number of i mmigrants Percent change Percent of U.S. immigrants State 1990 2000 2007 1990– 2000 2000– 2007 2000– 2007 on 10 year basis 1990 2000 2007 Total in U.S. 14,589,626 24,292,460 30,121,594 67 24 34 - - - Top Immigrant states - Total 10,778,377 16,770,500 19,897,845 56 19 27 73.9 69.0 66.1 California 4,933,152 7,101,428 8,083,580 44 14 20 33.8 29.2 26.8 New York 2,073,332 2,991,581 3,296,533 44 10 15 14.2 12.3 10.9 Texas 1,189,892 2,345,295 3,131,882 97 34 48 8.2 9.7 10.4 Florida 1,142,859 1,960,036 2,559,827 72 31 44 7.8 8.1 8.5 Illinois 720,573 1,214,660 1,429,508 69 18 25 4.9 5.0 4.7 New Jersey 718,569 1,157,500 1,396,515 61 21 29 4.9 4.8 4.6 Select high growth states - T otal 612,086 1,874,039 2,800,509 206 49 71 4.2 7.7 9.3 Arizona 199,473 507,084 784,769 154 55 78 1.4 2.1 2.6 Georgia 135,717 473,757 710,728 249 50 71 0.9 2.0 2.4 Colorado 107,796 288,427 391,970 168 36 51 0.7 1.2 1.3 North Carolina 88,311 348,962 505,699 295 45 64 0.6 1.4 1.7 Nevada 80,789 255,809 407,343 217 59 85 0.6 1.1 1.4 SOURCE : Author ’s calculations from IPUMS Decennial Census and American Community Survey data. N OTE : Includes only immigrants age d 18- 64. First, we note that the number of immigrants living in the United States increased by about 10 million from 1990 to 2000 and by about 6 million from 2000 to 2007 . A large fraction of the immigrant population lives in the historically popular destination states of California, New York, Texas, Florida, Illinois, and New Jersey. The United States and these six immigrant gateway states have seen increases in immigra nt population from 1990– 2007, but at a slower rate at the end of that period than at the beginning . In California, the immigrant population grew 44 percent from 1990 to 2000 but only 20 percent from 2000 to 2007 (on a 10-year basis) . Only eight states—Alaska, Hawaii, Louisiana, Maine, Montana, New York, Vermont, and West 14 Virginia —have lower immigrant population growth rates than California from 2000 to 2007. Since 1990, the share of immi grants choosing states with smaller immigrant populations and short histories of receiving immigrants has increased. In Arizona, Georgia, Colorado, North Carolina , and Nevada , immigrant populations grew at very high rates. 17 Changes in the immigrant population of California occur by various means. New arrival immigrants move into the state directly from other countries ; domestic (or internal) i mmigrants may also relocate from other states or to other states. In addition, some immigr ants may also leave the United States entirely. No nationally representative dataset is able to measure the outflow of immigrants from the United States accurately, so we do not consider the out- migration of immigra nts from the United States in this report. Although the growth rates in the se states declined from 1990 –2000 compared to 2000 –2007, as they did nationally, in four of them rates are nevertheless about double the rates of growth of the traditional states. T he share of immigrants choosing California fell by about 7 percent from 1990 to 2007 while the number choosing the new destination states rose about 5 percent over the same time frame. Indeed, t he decline in the percentage of new immigrants choosing the six gateway states is driven largely by the California’s decline . 18 Measuring Migration The 1990 and 2000 Census es and the American Community Surveys from 2005 to 2007 provide snapshot s of the population of states, counties, and MSAs . Looking at differences among these snapshots, we can estimate changes in population and demography. Because the Census asks the year of arrival for immigrants, we can estimate how much of the population change was due to new arrival immigrants. F rom this cross -sectional analysis we cannot estimate how much of the change was due to domest ic migration , but these datasets also allow study of migration at the individual level for a few distinct periods. In the 1990 and 2000 Census es, data were gathered on where respondents were living five years before. And in the American Community Surveys f or 2005 to 2007, data were gathered on where respondents were living one year before . From this, we can estimate the number of people that moved between states or MSAs over the 1985 –1990, 1995 –2000, and 2004 –2007 periods, allowing us to estimate the number of domestic in-migrants and out - migrants to a given area . Subtracting the number of out -migrants from the number of in- migrants gives the net internal migration of immigrants (or native-born) to the state or MSA. The components of net internal migration a re partly net figures already, as any intermediate moves within the period are not measured. Despite the drawbacks of measuring migration only within three- to five -year periods, the advantage of this Census data is the ability to break down migration sta tistics by individual 17 These new destination states are defined as being in the top 10 for growth and the top 25 for the number of immigrants. This definition gets around the problem of places with very few immigrants having very high growth rates because of the way growth rates are calculated rather than because there is a substantive change. 18 Various estimates suggest that up to 50 percent of immigrants to the United States eventually leave the country (Jasso and Rosenzweig (1982), Borjas and Bratsberg, (1996)). Some immigrants leave the country permanently, and some leave temporarily and return. The Census records the year of migration for immigrants, but does not ask whether this stay is the first or not. Our estimates do not consider out-migration and may be affected by immigrant s who have had multiple stays in the United States. Changes in death rates also affect the growth in California’s immigrant population, but given the relatively young average age of immigrants, changes in death rates are unlikely to affect overall estimates. 15 socioeconomic characteristics.19 We focus on persons of working age, because economic reasons tend to dominate migration decisions and because we are particularly interested in examining the effect of local labor market changes on migration patterns. (See T echnical Appendix A for background information on the determinants of location choice.) M ost analysis in this report is restricted to persons aged 18–64 who do not reside in institutions —the working age population. In parts of the analysis, we further restrict our attention to persons who report participation in the labor market (whether currently employed or unemployed). In particular, we are first interested in decomposing net internal migration into migration of native -born persons and migration of immigrants. As explained above, the stalling of growth of immigrant populations in historical immigrant destination states and MSAs may be caused by a decrease in new immigrant arrivals, an out -migration of previous immigrants, or both . In addition, changes in the concentration of immigrants, measured as a ratio of immigrants to total population, can also be caused by differential net internal migration of the native -born . So we proceed by first examining the number of new arrival immigrants and the net internal migration of the native- and foreign-born. We first examine whether the changes in immigrant population in California are driven by changes among newly arrived immigrants or among previous cohorts of immigrants. Table 3 shows that over these three periods, California’s growth in immigration population comes almost entirely from new arrivals. 20 Table 3 . Components of m igration in California Population change due to Change on a per-year basis California share of national new immigrant arrivals Net internal migration Net internal migration Years Net internal migration = Native born + Immigrant New immigrant arrivals Net internal migration = Native born + Immigrant New immigrant arrivals 1985–1990 120,714 81,333 39,381 942,795 24,143 16,267 7,876 188,559 34.9% 1995–2000 -407,162 -249,205 -157,957 906,935 -81,432 -49,841 -31,591 181,387 21.0 2004–2007 -388,374 -288,747 -99,627 698,836 -129,458 -96,249 -33,209 232,945 19.3 SOURCE : Author ’s calculations from IPUMS Decennial Census and American Community Survey data. N OTE : Includes immigrants age d 23– 64 as of 1990, 2000, or 2007, so as to count only persons of working age during the entire migration period. On an annual basis, the number of new immigrants to California was large and rela- tively steady during the periods 1985 –1990, 1995 –2000, and 2004 –2007. 21 19 We construct the 2004 –2007 migration period by linking one –year migration information across the three cross - sections of data. For this reason, we cannot track individual characteristics of 2004 –2007 migrants, only aggregate or group characteristics . However, California experienced an increase in the domestic out -migration of immigrants from 1995 to 2000 and 20 To view these migration trends relative to the state population, see T echnical Appendix Table A1. 21 For the comparison of the 2004– 2007 period to earlier periods, it is necessary to consider the population change on a per-year basis. Statistics from 1985– 1990 and 1995–2000 are divided by five and 2004– 2007 statistics are divided by three. Given the available data, this is the best comparison that we can do. However, the shorter migration period of 2004– 2007 may induce some bias. For example, we see a ri se in new immigrant arrivals in 2004–2007 on a per -year basis. It could be that the rate of new immigrants who arrived between 1995 –1998 was just as high but declined when the last two years of the period were measured. A comparison of new immigrants to Ca lifornia in the 2002–2007 and 1995– 2000 periods from the CPS suggests that indeed, the number of new immigrants fell when these full five-year migration periods are analyzed. Other limitations of the CPS data, however, prevent us from using it for the full analysis in this report. 16 from 2004 to 2007. During 1985–1990, more internal im migrants moved into the state than moved out, but during the latter two periods the opposite was true. Were it not for the increasing outflow of immigrants from California, the state would have seen a slightly larger growth in the immigrant population, because the inflow of new international immigrants remained roughly the same during the 1985–1990 and 1995 –2000 periods. Although on a per - year basis the inflow of new immigrants to California increase d from the 1995– 2000 period to the 2004– 2007 period , the percentage of all new immigrants who chose California continued to decline . In all periods, the number of newly arrived immigrants to California far outweigh ed the number of net migrants. So it is clear that the trends among new arrival immigrants drive the trends in overall immigrant population for the state. Comparing internal migration of immigrants to migration of native-born, we find that native -born net outmigration increased even more rapidly. 22 In addition, we find that California’s experience is unique among the immigrant gateway states , in that native-born outmigration mirrored that of immigrants, but did not accelerate as quickly . 23 In light of the declining popularity of California with new and previous immigrants, we next examine whether there is any evidence that the trend is driven by the choices of immigrants with certain socioeconomic characteristics. Who is it that California is no longer attracting? Are highly skilled immigrants choosing other states? Figure 2 shows the education distribution of California’s immigrants. In California and other traditional immig rant gateway states, the decline in new arrivals drive s most of the change in immigrant populations . In the new growth states, changes in immigrant populations are driven mostly by new arrivals as well, but are amplified by net increases due to internal mi gration of previous immigrants . In both group s of states, the trend among new immigrants is generally mirrored by the trend among domestic migrants. This marks a change from the recent past, where the flow of native-born domestic migrants tended to move in the opposite direction from the migration of immigrants. It also points to potential commonality in the reasons for migration between immigrants and the native-born . 22 This is a notable change from studies of migration over earlier periods that tend to find net domestic migrants move in the opposite direction from new immigrants (Frey and Liaw, 1998, and Bartel, 1989). However, Passel and Zim merman (2001) find a similar pattern for California during 1990–1995. 23 Technical Appendix Table A2 presents migration patterns for the six top immigrant states and the top new growth states. Technical Appendix Table A3 gives statistics for all states. 17 Figure 2. California’s Immigrant Education Distribution, 1990 –2007 SOURCE: Author ’s calculations from IPUMS Decennial Census and American Community Survey data . NOTE : Includes only immigrants age d 18–64 . The share of immigrants who had not completed high school fell from 39 percent in 1990 to 37 percent in 2000 and to 32 percent in 2007. Conversely, immigrants with a bachelor’s degree or higher increased from 17 percent in 1990 to 20 percent in 2000 and to 24 percent in 2007 . The shift in education distribution of immigrants can be explained primarily by changes in the education level of new immigrants. To show this, we decompose the components of migration in California along the education dimension. 18 Table 4 . Components of m igration in California by education Per year basis Net internal migration of New immigrant arrivals Percent of all new immigrants Years Education Net internal migration = Native born + Immigrant 1985–1990 < High school -4,562 -4,605 42 77,588 41 High school -6,272 -6,797 524 42,110 22 Some college 1,606 -520 2,127 32,179 17 College graduate 33,371 28,188 5,183 36,682 19 Total 24,143 16,267 7,876 188,559 1995–2000 < High school -25,806 -5,917 -19,889 61,315 34 High school -34,470 -24,071 -10,398 39,312 22 Some college -35,936 -31,016 -4,920 27,519 15 College graduate 14,779 11,163 3,615 53,242 29 Total -81,432 -49,841 -31,591 181,387 2004–2007 < High school -22,728 -6,147 -16,581 69,181 30 High school -43,509 -29,182 -14,327 51,427 22 Some college -47,928 -41,438 -6,490 30,149 13 College graduate -15,293 -19,482 4,189 82,188 35 Total -129,458 -96,249 -33,209 232,945 SOURCE : Author’s calculations from IPUMS Decennial Census and American Community Survey data . N OTES : Includes immigrants age d 23– 64 as of 1990, 2000, or 2007, so as to count only persons of working age over the entire migration period. Table 4 shows that there has been a net outflow of individuals with less than a high school diploma from California in all three periods . T his outflow is composed mostly of native- born p ersons from 1985 to 1990 but then largely of immigrants during the 1995 –2000 and 2004– 2007 periods . At the other end of the education spectrum, California saw a net increase in college -educated persons—both native -born and immigrant domestic migrants—during the 1985– 1990 and 1995 –2000 periods . Even with a net loss in college-educated native-born migrants during 2004 –2007, California still had a net increase in college -educated immigrants who moved within the United States from 2004 to 2007. Combining this in-migration of college - educated immigrants and out-migration of high school dropout immigrants produces an overall increase in education among immigrant domestic migrants to California . Newly arriving immigrants, who drive the trends for immigrants overall, are increasingly likely to have a college degree over these three time periods and decreasingly likely to have less than a high school diploma . This is the first time in recent California history that the proportion of new highly educated immigran ts exceeded that of less educated immigrants. California’s new immigrant arrivals appear to look different now in terms of education level . Changes in the educational composition of California’s immigrants may forecast changes in resource needs fo r the state, for example bilingual education . But education is only one characteristic of new immigrants. Understanding the new trends in immigrant settlement—and the implications for the state—requires a look at additional socioeconomic dimensions . 19 Changes in the Composition of California’s New Immigrant Arrivals As shown above, new immigrants to California have higher levels of education, on average, in 2007 than in 2000 or 1990. We next explore whether this trend is unique to Califor- nia , consider characteristics of new immigrants in addition to their education level s, describe the socioeconomic characteristics of new immigrants to California relative to new immigrants in other parts of the country , and examine how these characteristics have changed over time. Historically, immigrants have not been spread out evenly across the United States, but rather have t ended to cluster in a relatively small number of cities and states and often locate near other immigrants with similar socioeconomic character istics. For example, both Florida and California are popular immigrant destination states, but Florida is much more likely to attract Cuban immigrants than California is, and California is a more likely destination for Mexican immigrants than Cuban ones . O ver time, however, this clustering has declined, an d markedly so for California . We use statistical models that relate an immigrant’s individual characteristics to the choice of living in California relative to other states in each of the years 1990, 2000, and 2007. 24 These models reveal how immigrants’ socioeconomic characteristics, on average, are related to their choice of living in California rather than other states . For each year, we estimate the likelihood of new immigrants choosing to live in the state given a large set of characteristics, such as age, education, ethnicity 25, gender, marital status, English fluency, employment, wage level, industry of employment, housing costs, and homeownership. 26 The models also show that although new immigrants to California look very different from those moving to other states, in socioeconomic terms this disparity has declined since 1990. 27 24 These are simple linear probability models. Probit estimates are similar. Figure 3 shows the estimated probability that an immigrant chose to live in California in 1990, 2000, or 2007, and shows how that probability changes across immigrants’ characteristics. (The chart displays only a handful of characteristics included in the statistical model; for full results see Technical Appendix Table A4. ) A typical immigrant wa s about 37 percent more likely to choose California than other states in 1990 (“Average” column ). This probability decline d to about 18 percent in 2007. If an immigrant had less than a high school degree, he or she was even more likely to choose the state (comparing the column “Less than High School” to the horizontal line ). While less educated new immigrants a re more likely to live in California than other states, this clustering of less educated new immigrants in California has declined over time. An immigrant with less th an a high school degree wa s about 40 percent more likely to choose California in 1990 . By 2007, this wa s down to 23 percent. Conversely, college - educated new immigrants were less likely to choose California, but the difference decreased over time (“College Degree” column) . 25 Throughout this report we define ethnicity variables based on self -reported “race” categories given in the Census samples. They are as follows: Latino, non -Latino White, non-Latino Black, and Asian. All others are included in a final “other” category. 26 Technical Appendix B provides more detail on some of these variables. 27 See Technical Appendix Table A4 for detailed regression results. 20 Figure 3 . Predicted Probability of Living in California for New Immigrants with Selected Characteristics SOURCE: Author’s calculations from IPUMS Decennial Census and American Community Survey data. NOTES: Includes only imm igrants aged 18-64 with 0-5 years in the United States. Bars represent predicted probability from linear probability model described in text. “Average” is the predicted probability at the average value of explanatory continuous variables and taking account of the distribution across categorical va riables. All other bars use these average values except for the characteristic listed . By comparing the bar to the horizontal lines for 1990 or for 2000 and 2007, one can estimate the percentage point difference in predicted probability for each factor compared to the average predicted probability in that year . Another notable feature of the immigrant clustering in California is its ethnic dimension. A new Latino or Asian immigrant wa s more likely to live in California than other states in 1990, Latinos roughly 20 percent more likely , and Asians 1 4 percent more likely. But this overrepresentation has decline d over time. By 2007, a Latino new immigrant wa s only 5 percent more likely to live in California a nd an Asian new immigrant 13 percent. To see this graphically, we compare immigrants who are either Latino or Asian to all immigrants. The probability of choosing California was still high, but declined markedly over time, especially fo r Latino s. Compared to average, a Latino immigrant was about 7 percent more likely to choose California than other states in 1990. By 2007, a Latino immigrant was no more or less likely than an average immigrant to choose California . This trend for Latino immigrants in particular stands in contrast to the state’s history of attracting a large percentage of new Latino immigrants. New immigrants to California cluster not only by education and ethnicity, but also along some economic dimensions. In 1990, new immigrants to California were much more likely to be 0 5 10 15 20 25 30 35 40 45 50 Av er a geLa t in oAsianLess than High School College Degr ee Co n s t ruct ion Industry Ma n u fa cturing Industry Predicted Probability New Immigrant with Average Characteristics Chooses Calif ornia 199020002007 36 20 19.8 Probability for Avg Immigrant in 1990 Probability for Avg Immigrant in 2000 and 2007 21 employed in agriculture, construction, manufacturing, and in some service industries than were new immigrants to other states. Each of these industries, except manufacturing, employed a larger fractio n of the labor force in California than in other states (see Te chnical Appendix C). The predominance of these industries is likely a draw for new immigrants, but the industry composition of employment in the state is also affected by the employment of new immigrants. The likelihood of new immigrants employed in construction and manufacturing industries to choose California over other states in 1990 is shown in the height of the darkest bars in the Figure 3 Construction and Manufacturing columns. For new imm igrants in construction, by 2000 , they we re more likely to live outside California (shown in the fall of predicted probability in the Figure 3 Construction column). Note that this coincides with an overall drop in the share of Californians working in const ruction. In manufacturing, by 2007, new immigrants in California were about as likely to work in this industry as we re immigrants in other states. We see similar declines in new immigrants in California working in business services, despite an increase in the percent of the California economy employed in business services (Technical Appendix C) . Thus, changes in California’s industry composition cannot be the only draw for new immigrants to the state. In su mmary, in 1990, 2000, and 2007, we find significant clustering of new immigrants in California along social and economic dimensions . New immigrants with less than a high school degree or Latino new immigrants are more likely to live in California than other states, and those with a college degree are less likely . However, the degree of this clustering has declined over time. The changes in characteristics of the new immigrants to California along ethnic and education dimensions in particular stand in contrast to a long historical trend of immigration to the state. 22 Choices within California Although the large inflow of new immigrants to California continues, more new immigrants to the United States are choosing to live in other states. Similarly, it appears that as immigrant populations grow in new destination states, they are also growing in new areas of this state. Nationally, new immigrants who chose California were like new immigrants to other states along socioeconomic dimensions after 2000 tha n they were before 2000. In this section we examine whether the same can be said for trends among new immigrant s to different areas wit hin California . Within California, Los Angeles County receives the most new immigrants but also has the largest outflow of immigrants to other counties or states. Table 6 shows the components of population change in some of the major counties in the state. 28 F our counties always rank at the top based on immigrant population since 1990: Los Angeles, Orange, San Diego, and Santa Clara . In all of these historically immigrant -rich counties except Santa Clara, the number of new immigrants fell from the 1985 –1990 period to the 1995 –2000 period and rebounded in the 2004– 2007 period, on a per -year basis. 29 Most of California’s gateway counties experienced a net outmigration due to internal relocation since 1985. In all gateway counties except Santa Clara, net outmigration accelerated from 1985 to 2007. Over the latter two migration periods we measure, the outmigration is not just of native-born but also of immigrants. To some extent this mirrors the slowing of new immigrant arrivals to these areas. This finding runs counter to research on earlier periods arguing that primarily native -born persons respond to immigrant influxes by relocating. However, in all the gateway counties, the number of new arrivals was vastly larger than the increase from internal migration . This is consistent with the nationwide trend that in gateway areas, new immigrant arrivals drive the change in immigrant populat ion. 30 28 See Technical Appendix Table A6 for sta tistics on all counties in California. It suggests that the factors underlying the changes in settlement patterns for immigrants are common factors to the location decisions for native-born individuals . 29 For the comparison of 2004 –2007 to earlier periods, it is necessary to consider the population change on a per year basis; 1985– 1990 and 1995– 2000 changes are divided by 5, and 2004-2007 divided by 3. Given the available data, this is the best comparison that we can do. However, the shorter migration period of 2004 –2007 may induce some bias. 30 Frey and Liaw (1998), for example. 23 Table 6. Components of migration in Calif ornia counties 1985–1990 1995–2000 2004–2007 Net internal migration of: Net internal migration of: Net internal migration of: County Net internal migration = Native born + Immigrant New immigrant arrivals Net internal migration = Native born + Immigrant New immigrant arrivals Net internal migration = Native born + Immigrant New immigrant arrivals Top immigrant destinations Los Angeles -283,810 -225,756 -58,054 440,861 -306,646 -149,416 -157,230 319,314 -294,952 -159,689 -135,263 220,417 Orange -23,968 -32,454 8,486 92,385 -31,250 -19,955 -11,295 84,720 -57,614 -36,616 -20,998 63,990 San Diego 43,327 31,107 12,220 62,647 -33,456 -27,578 -5,878 58,845 -68,085 -60,240 -7,845 57,687 Santa Clara -41,565 -42,747 1,182 59,234 -51,025 -50,587 -438 95,280 -16,397 -20,973 4,576 67,659 Top growing immigrant destinations Alameda -19,763 -25,915 6,152 36,257 -4,975 -22,136 17,161 59,450 -13,265 -10,369 -2,896 42,720 San Bernardino 119,517 92,104 27,413 24,103 -2,358 -14,525 12,167 23,161 23,546 -513 24,059 21,831 Riverside 131,698 103,769 27,929 21,067 48,212 36,798 11,414 24,017 103,959 53,288 50,671 25,142 Sacramento 41,928 36,247 5,681 12,003 5,547 1,086 4,461 23,903 1,327 1,443 -116 23,294 Kern 11,636 10,315 1,321 8,089 -12,934 -10,037 -2,897 8,244 16,776 9,137 7,639 8,628 SOURCE: Author’s calculations from IPUMS Decennial Census and American Community Survey data. NOTES: Includes immigrants age d 23 –64 as of 1990, 2000, or 2007, so as to count only persons of working age over the entire migration period. 24 Alameda, San Bernardino , Riverside, Kern and Sacramento Counties are fastest growing in the state in terms of immigrant population. 31 The re are two distinct patterns of growth , shown in the bottom panel of Table 6 . 32 In Alameda and Sacramento counties, growth in immigration comes predominantly from growth in new immigrant arrivals. These areas might thus be deemed new immigrant gateways. In San Bernardino and Riverside C ounties, however, more of the growth comes from internal migration of older arrival immigrants. During the 1985– 1990 and 2004 –2007 periods , more immigrants moved into San Bernardino and Riverside from domestic locations than from other c ountries. These counties are adjacent to Los Angeles County, and th e previous inflow of immigrants was due primarily to immigrants relocating from Los Angeles. 33 Just as immigration increases to new destination states have spurred legislation , so have the increases in local areas within California with little history of immigration . For example, San Bernardino County ha s made national news with a proposal for restrictive ordinances to regu - late immigration. These findings indicate that Los Angeles continues to function as a gateway for immigrants who subsequently move elsewhere, especially to nearby inland counties . Kern C ount y’s migration pattern lies in between the two, with little growth in new immigrant arrivals, but a significant inflow of older arrival immigrants only from 2004–2007. 34 Figure 4 shows the education distribution of immigrants in six state MSAs . In 1990, 42 percent of the immigrant population of the Los Angeles MSA had not fi nished high school, but by 2000 that ratio was down to 39 percent and to only 32 percent by 2007. Many MSAs in the state experienced increases in average education but none at as high a rate as Los Angeles. Similarly, most MSAs in the state—even new growth areas like Sacramento —experienced an increase in the percent of immigrants who had a college degree. Not surprisingly , given the nature of their industries, the San Francisco and San Jose MSA s have the highest fraction of immigrants with college degrees, 35 percent and 45 percent in 2007, respectively , up from 26 percent and 29 percent in 1990. Such action could be sparked simply by overall growth of the immigrant population in areas with little history of incorporating immigrants, but could also be related to particular socioeconomic changes driven by new immigrant arrivals. 35 Unlike new -growth states across the country, many of which experienced an increase in the fraction of immigrants with less than a high school diploma, new-growth MSAs in Cali- fornia saw declines in the share of these immigrants, similar to the trend for the state overall. 31 As noted previously for high growth states, we define high immigrant growth counties as the top 10 in terms of growth over 1990 –2000 and at least in the top 20 in terms of number of immigrants as of 1990. 32 To view these migration statistics as rates per number of residents, so as to compare the changes across areas of different size, see T echnical Appendix Table A7. 33 Results available upon request. 34 Chang, Cindy, “California City Council Rejects Anti -Immigration Legislation,” New York Times, May 16, 2006. National Public Radio, Morning Edition, “California Town Aims to Bar Illegal Immigrants from Renting,” May 12, 2006, accessed 1/22/09 at: http://www.npr.org/templates/story/story.php?storyId=5400392. 35 Santa Clara County has a high proportion of its workforce employed in manufacturing relative to other counties and San Francisco ci ty-county has a high proportion employed in retail trade, finance, and professional services. See T echnical Appendix C. 25 Figure 4. California MSAs Immigrant Education Distribution SOURCE : Authors’ calculations from IPUMS Decennial Census and American Community Survey data. NOTE : Includes only immigrants age d 18– 64. As in the previous section, we examine changes in socioeconomic characteristics of immigrants across California in a multidimensional manner , rather than by examining education levels alone . Earlier , we used statistical models to describe the characteristics of new immigrants who choose California relative to those choosing other states. Since Los Angeles County accounts for about half of all immigrants in the state, and because the decline in immigration to Los Angeles drives the decline in the overall state immigrant population, we will examine the choice of Los Angeles over all other areas in the state. As above, we estimate 42% 24% 19% 39% 24% 32% 27% 18% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates Los Angeles-Long Beach, CA 23% 31% 20% 35% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates San Francisco-Oakland-Vallejo, CA199020002007 44% 24% 11% 43% 26% 19% 12% 37% 30% 19% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates Riverside-San Bernadino, CA 24% 20% 39% 17% 45% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates San Jose, CA 34% 24% 17% 31% 24% 28% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates San Diego, CA 28% 19% 27% 21% 26% 0% 10% 20% 30% 40% 50% Percent High School DropoutsPercent High School GraduatesPercent With Some CollegePercent College Graduates Sacramento, CA 26 how new immigrants’ characteristics are related to the likelihood of choosing to live in Los Angeles compared to that of living elsewhere in the state. The models corroborate the findings on education distribution discussed above. 36 When we examined California overall, we found evidence that the industry of employ - ment for immigrants who choose the state has changed from 1990 to 2007 . In particular, the likelihood of immigrants who work in construction and manufacturing choosing California declined. However, when looking within the state at the choice of Los Angeles ove r other areas, we find the strongest clustering of new immigrants in Los Angeles in service industries. The draw of the entertainment and hospitality service industry in the Los Angeles MSA increases over the period of study . Similarly, the likelihood of n ew immigrants in the finance, insurance, and real estate industry in Los Angeles has increased sharply relative to other areas in the state. New immigrants who work in construction are less likely to live in Los Angeles in 2000 and 2007, and we also see de clines among immigrants who work in the manufacturing industry. The decline in new immigrants who work in manufacturing is consistent with the overall decline in the percent of the Los Angeles workforce in manufacturing from 1990 to 2007, but there was no similar decline in the overall percent of the Los Angeles workforce employed in service industries or construction. Latino new immigrants were 11 percent more likely to live in the Los Angeles MSA in 1990, but only 6 percent more likely in 2007 . New immigrants who decide to live in Los Angeles are more likely to lack a high school diploma , but that is less so over time. New immigrants in Los Angeles earn less in wages than other new immigrants in the stat e, after controlling for indiv- idual characteristics related to earnings potential , moving from about 1 percent less in 1990 to 6 percent less in 2000 and 2007. 37 Immigrants’ location decision s are not only dependent on their characteristics but also on the conditions of the different geographic areas the y have to choose from. In the next section, we combine the analysis of the characteristics of new immigrants to California and the charac - teristics of the locations they choose from in order to uncover some of the factors behind these immigrants’ settlement decisions. 36 See Technical Appendix Table A8 for regression results. As before, we execute a linear probability model of living in the Los Angeles MSA relative to elsewhere in the state. 37 See Technical Appendix C for details. 27 Have New Immigrants’ Preferences Changed? Immigrants choose specific destinations for a host of reasons . R esearch has shown the primary reason behind immigrant location choice is social ties, with economic factors a distan t second . However, in light of new settlement patterns and changes in the characteristics of immigrants choosing California, the importance of these factors may have changed . Immigrants to the United States and residents who consider relocating within the coun- try weigh many factors . There are a number of theories of migration and location choice, which give varying weight to determinants that fall roughly into the economic, social, and institutional categories. 38 Previous research finds that economic costs and benefits are also significant factors in immigrant location choice, but are second to social fact ors. Among measures used to analyze the economic draw of a locale are jo b opportunity, wage level, employment rate, and housing prices. Individual-level characteristics such as education, work skills, and occupational status also factor into location choice since they are strongly tied to economic opportunity . The relative importance of these economic factors is difficult to gauge. For example, the draw of high wages often coincides with high housing prices, which otherwise may be undesirable. Thus, a number of economic factors may work together in complex ways. The distance to one’s home country is related to the cost of a given location, not simply economic cost but also psychic costs. This factor is thus interpreted as both an economic and social factor and will be included in the statistical models here. Social factors for migration, in particula r the concentration of co-ethnics in a local- ity, have been found to be the strongest predictor of immigrant location choice. This measure is very broad, but is used as a proxy for social ties to a particular location, for job opportunities arising from social connections, and for other associated informational or social benefits . Some institutional factors that are postulated to affect migration include state welfare benefit generosity, border crossing locations, and broad changes in Unites States immigration policy . Analysis of these factors is limited by other economic and social changes that coincide; however , we are able to include a measure of state welfare benefit generosity . The model estimates the probability that a new immigrant to the United States chooses an MSA based on its social, economic, and institutional characteristics . To do this parsimoni - ously, I use a multinomial framework, where individuals choose among MSAs as a function of individual and MSA characteristics. 39 We focus on Latino and Asian immigrants only in these models; further detail on immi - grant ethnicity cannot be estimated due to limited sample size. The statistical models require a large sample of individuals in order to estimate the effects we are interested in accu rately. By To incorporate differences across immigrants, the models are estimated separately by ethnicity. To examine changes in the importance of the factors over time, the models are also estimated separately by year. Details on some of the more technical aspects of this location choice model and full re gression statistics are available in Technical Appendix D. 38 For further detail on the theory and findings of research on immigration location choice, see Technical Appendix A. 39 In particular, I use the conditional logit model of McFadden (1973). The probability P(i,j) that individual i chooses location j is a function of MSA characteristics Z and individual characteristics X. 28 looking at only these two gro ups, we cover roughly 70 percent of the new immigrant popula - tion. New Latino immigrants make up 43 percent of new immigrants from 2002 to 2007 and new Asian immigrants constitute 30 percent. We find that the trends in location choice vary between Latinos and Asian immigrants, so the results for each group are discussed separately . Have Latino Immigrant Preferences Changed ? First we examine the determinants of location choice among new Latino immigrants. I n Figure 5, w e see that in 1990, having co -ethnics in an MSA was a very strong draw for new immigrants. The percent of the MSA population of the same ethnicity is used to measure the presence o f co-ethnics in these models. A 10-point higher fraction of Latino immigrants in an MSA was related to a 25 - percent higher chance that new Latino immigrants would choose that MSA . Similarly, increasing distance from the home country wa s a deterrent. However, these factors have decline d in their importance over time. 40 We find little evidence that new Latino immigrants choose M SAs within states that have more generous welfare benefits. An additional $100 per year in TANF or AFDC benefits in an area was related to a 2.5 -percent higher chance that a new immigrant would choose t hat location in 1990, and to only a 1-percent chance in 2000 or 2007. By 2000, new Latino immigrants were not strongly influenced by proximity to others of the same ethnicity . In fact, a 10-point higher share of Latino immigrants in an MSA ma de it slightly less likely that new Latino immigrants would locate there. This marks a turnaround in the determinant which has historically predominated immigrants’ location decisions . Social factors are still very important in the location choice of new Latino immigrants —overall, a large percent of new Latino immigrants choose Los Angeles, for example —but in comparing two areas with different concentrations of co-ethnics, the relevance of this factor ha s fad ed. 41 40 To view the summary statistics on these explanatory variables, see Technical Appendix Table D2. The increase in average percent of Latinos across MSAs over time does not explain the fall in the coefficient estimates. Also, average distance to home country has declined, so if anything we would expect the coefficient on distance to home country to in crease over time. This relationship is difficult to identify, since state welfare generosity is related to other economic and social conditions that might attract immigrants. 41 An $100 per year increase is about a 1-percent increase at the average benefit level across states. See Technical Appendix Table D2. 29 Figure 5. Importance of Factors in Latino New Immigrant Location Choice SOURCE: Authors’ calculations from IPUMS Decennial Census and American Community Survey data. NOTE: Includes immigrants aged 23-64 as of 1990, 2000, or 2007, so as to count only persons of working age over the entire migration period, and immigrants with 0 -5 years residence in the United States. Bars represent estimated effect on probability of choosing an MSA given the changes in covariates listed on the x -axis. Only statistically significant results at the 1 -percent level are shown . Full model statistics are presented in Technical Appendix Table D1. Looking at economic determinants of location choice among new Latino immigrants, we find evidence that new immigrants are drawn to MSAs that have jobs available —the unemploy - ment rate was an important indicator over all periods . The higher the unemployment rate, the less likely a new Latino immigrant is to choose an MSA . For a 1 -percent higher unemployment rate in an MSA, a Latino immigrant in 1990 wa s 15 percent less likely to choose that location . By 2007, thi s was down to about 6.5 percent , al though still statistically significant . New Latino immigrants are drawn to MSAs with high average wages, but only in 1990. As noted, it is difficult to predict the effect of higher wages or higher housing prices on location decisions . We might expect that immigrants would choose an MSA with a low cost of housing, holding everything else constant. However, labor market conditions, especially wages, are related to housing market conditions and are a very strong draw . T he correlation of these variables makes it hard to separately identify the draw of low-cost housing , and this probably explains the relatively small effects. 42 42 Excluding housing or wages from the model has little or no effect on other estimated coefficients. -20 -15 -10 -5 0 5 10 15 20 25 30 10 pct pt higher concentratio n of co-e thnic s 100,000 more people$100 higher weekly wage1 pct pt higher une mp. ra te$100 mor e / mon t h max welfare benefit 100 mor e mile s to home c ou ntry$10,000 higher hou sing c os t inde x Increase or Decrease in Probability of Choosing MSA Change in MSA Characteristic of Interest 1985-1990 Latino Immigra nt Arriva ls in 1990 1995-2000 Latino Immigra nt Arriva ls in 2000 2002-2007 Latino Immigra nt Arriva ls in 2007 30 Have Asian Immigrants’ Preferences Changed ? In contrast to the results for new Latino immigrants, changes in the draw of social fac - tors for new Asian immigrants are mixed . Over time, n ew Asian immigrants are less likely to choose a destination due to proximity to their home countr ies. Because the average distance for an Asian immigrant to his or her home country is f arther than for a Latino immigrant, the decrease in probability due to distance to h ome country is much smaller. Asian immigrants are increasingly likely to choose an MSA because of the presence of other Asian immigrants there. Here agai n, we find no evidence that Asian immigrants choose their destinations be- cause of generosity of welfare programs. The change in probability of choosing an MSA because the area offers $100 more in welfare benefits is estimated at zero or nearly zero in all years . Economic factors are a similarly strong draw for new Asian immigrants. Asian immi- grants were 20 percent more likely to choose an MSA because it ha d $100 higher weekly wages, all else equal, in 2007 . This estimate for Asians is stronger than for Latino immigrants, where statistically significant effects of higher MSA wages were found only in 1990 . New Asian immigrants were strongly deterred from choosing an MSA with higher unemployment rates. 43 Last, the difficulty in estimating the effect of housing prices on immigrant location choice is evident here as well . It is unclea r from the estimates whether new Asian immigrants are consistently drawn to MSAs with higher housing prices or lower ones. We expect that this is likely driven by the relationship of local area wages and housing prices. The size of the estimates is consistent with the size estimated for Latino immigrants, although both show evidence that this factor has declined somewhat in importance over time. 43 Additional testing reveals that the deterrent effect of the unemployment rate increases as the rate increases. 31 Figure 6. Importance of Factors in Asian New Immigrant Location Choice SOURCE: Authors’ calculations from IPUMS Decennial Census and American Community Survey data. NOTE: Includes immigrants aged 23-64 as of 1990, 2000, or 2007, so as to count only persons of working age over the entire m igration period, and immigrants with 0 -5 years residence in the United States. Bars represent estimated effect on probability of choosing an MSA given the changes in covariates listed on the x -axis. Only statistically significant results at the 1 -percent level are shown . Full model statistics are presented in Technical Appendix Table D1. Taken together, this simple model leads to some striking conclusions. The first is the decline in social factors as a driver of location choice . We still observe a great amount of clustering of immigrants along ethnic and home country lines, but the evidence here suggests this factor may play less of a role over time. Indeed, it reflects the dispersion of immigrants to new destination areas. Second, we find no proof that w elfare programs are a prominent factor in the location choice of new immigrants. On the other hand, there is evidence that economic factors have played a strong and consistent role in the location decision s of new immigrants. For both new Latino and Asian immigrants, higher average wage has a positive effect on the probability of choosing an MSA and unemployment rate has a negative one. These changes may be interpreted in different ways with regard to the assimilation of new immigrants. New trends in immig rant location choices to new destinations that have little history of immigration present challenges to local areas. The social adaption of immigrants may be slowed because they are less likely to choose areas simply because of the availability of immigrant networks that help to facilitate assimilation . However, the long-term success and broad assimilation of immigrants is also related to economic progress. The econo mic oppor- -20 -10 0 10 20 30 40 10 pct pt higher concentration of co-ethnics 100,000 more people$100 higher weekly wage1 pct pt higher unemp. rate$100 more/mont h max welfare benefit 100 more miles to home country$10,000 higher housing cost index Increase or Decrease in Probability of Choosing MSA Change in MSA Characteristic of Interest 1985-1990 Asian I mmigrant Arrivals in 1990 1995-2000 Asian I mmigrant Arrivals in 2000 2002-2007 Asian I mmigrant Arrivals in 2007 32 tu nity that new destinations present may spur faster economic assimilation and long -term success of new immigrants. Economic motivations for location choice may mitigate the challenges to new destina - tion areas . To the extent that economic opportunity leads to economic success, new immigrants motivated increasingly by economic opportunity are more likely to contribute to, rather than tax the system. In California, the recent decline in the popularity among new immigrants of the state itself and of the Los Angeles region in particular may be understood better in light of the preferences of new immigrants. Proximity to Mexico and to large Latino and Asian immigrant populations have historically driven continued flow of new immigrants to the state. The decline in importance of these factors explains a good portion of the decline in the state’s popularity. But we also find that economic opportunities are consistently a strong factor behind immigrant location choice . Additional r esults suggest that the location decisions of highly educated Latino and Asian new immigrants are even more strongly related to economic conditions in MSAs. 44 44 See Technical Appendix Figures D1 and D 2. Thus , the state’s ability to attract a highly skilled immigrants for its workforce is linked to economic conditions in other places. 33 Figure 7. Trends in Unemployment, Wages, and Housing Prices in California and the United States With the decline of social factors as a draw for immigrants, the state’s economic condition has increasing importance for drawing immigrants, and drawing highly educated immigrants in particular. The current economic crisis will have an impact on the numbe r and characteristics of immigrants who choose to come to the United States and who choose to live in California . The recent upturn in unemployment rates is seen in the third panel of Figure 7 . The California and Los Angeles unemployment rates have increased more quickly in the last year . To the extent that the economic decline is more pronounced in California than in other areas, $0 $50,000 $100,000 $150,000 $200,000 $250,000 $300,000 $350,000 $400,000 $450,000 $500,000 198519871989199119931995199719992001200320052007 Median House Price Median MSA Home Prices California MSAs Los Angeles-Long Beach MSA All US MSAs SOURCES: OFHEO Housing Price Index, US Census Bureau. NOTES: Median house prices respresent average across MSAs; 1999 dollars. $0 $5,000 $10,000 $15,000 $20,000 $25,000 $30,000 $35,000 $40,000 $45,000 $50,000 198519871989199119931995199719992001200320052007 Average Annual Wage Annual Wages for All Workers and New Immigrants All California WorkersAll US WorkersNew California ImmigrantsNew US Immigrants SOURCE: Current Population Survey (US Census, Bureau of Labor Statistics). NOTES: 18 -64 year-oldm ale s o nly; immigration status not available prior to 1994; 1999 dollars. 0% 2% 4% 6% 8% 10% 12% 198519871989199119931995199719992001200320052007 Average Annual Unemployment Rate Average Annual Unemployment Rates California Los Angeles-Long Beach MSA US SOURCE:Bureau of Labor Statistics. NOTE: Average annual unemployment rates are not seasonally adjusted. 34 the state will likely be disadvantaged in attracting immigrants in the future. However, the first two panels of Figure 7 , showing median housing prices and wage rates, are not definitive in predicting how California’s draw for new immigrants might change. Even with the economic downturn, California has a higher average wage than the rest of the United States, a draw , and the state’s markedly higher housing prices , a deterrent, have started to decrease. 35 Conclusion Although California is still home to more immigrants than any other state, its popularity for immigrants began to wane for the first time after 1990 . Calif ornia ’s 44-percent immigrant growth rate from 1990 to 2000 was far exceeded by growth rates of more than 200 percent in states such as Georgia and North Carolina . Within California, the number of immigrants in Los Angeles County increased an average of ju st 1.9 percent per year during the 1990 –2007 period , compared to rates as high as 12.6 percent per year in Riverside County and 10.5 percent in Kern County. In this report, we find that most of the change is driven by the changing location choice of new arrival immigrants rather than those of previous immigrants. Within California there is also a significant shift of immigrants to new destinations within the state, although the traditional immigrant gateway cities of Los Angeles and San Francisco still serve as jumping off points for new immigrants. Thus, programs to facilitate new immigrant assimilation in gateway areas are likely to continue to have a large population to serve. The trend towards settlement in new destinations is also driven in many areas by the internal migration of native -born persons, suggesting that determinants of location choice are increasingly common to both groups . Some researchers have shown that in the recent past, native -born persons tended to leave areas with high in -migration of immigrants. We find a reversal of this trend . In addition , our findings suggest that immigrant inflows are not responsible for initiating native -born outflows. Thus , it may be that the level of competition between native-born and immigrants has been o verstated. New Latino immigrant arrivals to the United States were much less likely to choose to live in California in 2007 than they were in 1990; new immigrants employed in construction, manufacturing, and some service industries also prefer other states. New immigrants choosing California over other states were slightly more educated in 2007 than in 1990 . We find that changes in the composition of new immigrants are related to the factors by which they make their decisions on wh ere to settle. Social factors , historical ly the principal determinant , have waned in importance for Latino immigrants in particular . New Latino immigrants are much less likely to choose cities because of the presence of co -ethnics. These social factors are still the primary explanation for location choice, but economic factors are a strong second . This explains the decision of many immigrants to live in new destinations that h ave a less established immigrant social network but have growing economic opportunities. The established immigrant networks in California have less attraction for the newest immigrants to the United States, and this explains most of the decline in the state’s popularity. The waning importance of social factors in new immigrants’ locati on decisions may signal the assimilation of immigrants . For those concerned with the integration of immigrants, the decline in clustering of immigrants along social lines and the increasing i mportance of economic factors may be good news . On the other hand, in areas with few immigrants and little experience incorporating immigrants into social and economic life, the settlement of immigrants away from co -ethnics raises new challenges. 36 References Bartel, Anne , “Where Do the New U.S . Immigrants Live?” Journal of Labor Economics, V ol . 7, No. 4 , 1989. Bohn, Sarah, “Immigration and Wages in the U.S. Labor Market ,” University of Maryland Dissertation, UMI Number 3283414 , 2007. Borjas, George , “The Labor Demand Curve Is Downward Sloping : Reexamining the Impact of Immigration on the Labor Market, ” Quarterly Journal of Economics, V ol . 118, No. 4 , 2003. Borjas, George , “Immigration and Welfare Magnets,” Journal of Labor Economics, Vol. 17, No. 4 , 1999. Borjas, George , and Bernt Bratsberg , “Who Leaves? The Outmigration of the Foreign -Born,” The Review of Economics and Statistics, V ol . 78, No. 1 , 1996 . Frey, William , and Kao -Lee Liaw , “The Impact of Recent Immigration on Population Redistribution Within the United States,” in James P. Smith a nd Barry Edmonston, eds., The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration, National Academy Press, Washington D .C ., 1998. Gibson, Campbell, and Emily Lennon, “Historical Census Statistics on the Foreign -Born Population of the United States 1850 -1990, ”U.S. Census Bureau Population Division Working Paper No. 29, 1999. Hopkins, Daniel, “Threatening Changes: Explaining Where and When I mmigrants Provoke Local Opposition,” Center for the Study of American Politics, Yale University , 2007. Jaeger, David , “Local Labor Markets, Admission Categories, and Immigrant Location Choice,” IZA Discussion Paper, 2000. Jasso, Guillermina , and Mark Rosenzweig , “Estimating the Emigrati on Rates of Legal Immigrants Using Administrative Survey Data: The 1971 Cohort of Immigrants to the United States, ” Demography , Vol. 19, No. 279–290, 1982. Kaushal, Neeraj , “New Immigrants’ Location Choices: Magnets without Welfare,” Journal of Labor Economics, Vol. 2 3, No. 1, 2005 . Lewis, Ethan, “Immigration, Skill Mix, and the Choice of Technique,” Federal Reserve Bank of Philadelphia Working Paper No. 05- 8, 2005. McFadden, D aniel, “Conditional Logit Analyses of Qualitative Choice Behavior ,” in P. Larembka, ed., Frontiers of Econometrics, Academic Press, New York, 1973. Massey, Douglas, Jorge Durand, and Nolan Malone , Beyond Smoke and Mirrors: Mexican Immigration in an Era of Economic Integration, Russell Sage Foundation, New York, 2002. Massey, Douglas, ed. , New Faces in New Places: The C hanging Geography of American Immigration, Russell Sage Foundation , New York, 2008 . 37 National Conference of State Legislatures, “2007 Enacted State Legislation Related to Immigrants and Immigration”, available at www.ncsl.org/programs/immig/2007immigrationfinal.htm, Jan. 31 2008 . National Conference of State Legislatures, “State Laws Related to Immigrants and Immigration: January 1 – June 30, 2 008”, available at www.ncsl.org/programs/immig/immigreportapril2008.htm, July 28 , 2008. National Conference of State Legislatures, “State Laws Related to Immigrants and Immigration: January 1 – June 30, 2008”, available at www.ncsl.org/programs/immig/immigreportjuly2008.htm, July 24 , 2008. National Conference of State Legislatures, “State Laws Related to Immigran ts and Immigration in 2008”, available at www.ncsl.org/programs/immig/2008StateLegislationImmigration.htm, January 27 2009. Passel, Jeffrey and Wendy Zimmerman , “Are Immig rants Leaving California? Settlement Patterns of Immigrants in the Late 1990s, ” Urban Institute, 2001. Peri, Giovanni , “How Immigrants Affect California Wages and Employment, ” Public Policy Institute of California , San Francisco , 2007. Ramakrishnan, S. K arthick, and Paul Lewis , “Immigrants and Local Governance: The View from City Hall ,” Public Policy Institute of California , San Francisco , 2005. Reed, Deborah , “California’s Future Workforce: Will There Be Enough College Graduates ?” Public Policy Institute of California , San Francisco , 2008. Ruggles, Steve, Matthew Sobek, Trent Alexander, Catherine A. Fitch, Ronald Goeken, Patricia Kelly Hall, Miriam King, and Chad Ronnander , Integrated Public Use Microdata Series: Version 3.0 [Machine -readable database] . Minneapolis : Minnesota Population Center [producer and distributor] , 2004. Singer, Audrey , “The Rise of New Immigrant Gateways,” The Brookings Institution Center on Urban and Metropolitan Policy, Washington, D.C., 2004. Singer, Audrey, Susan Hardwick, and Caroline Brettell, eds. , Twenty -First Century Gateways: Immigrant Incorporation in Suburban America, Brookings Institution Press, Washington, D.C., 2008 . Singer, Audrey , and Robert o Suro , “Latino Growth in Metropolitan America: Changing Patterns, New Loc ations,” The Brookings Institution , Metropolitan Policy Program, 2002. U.S. Department of Homeland Security, Task Force on New Americans, “ Building an Americanization Movement for the Twenty -first Century: A Report to the President of the United States fro m the Task Force on New Americans, ” Washington, D.C., 2008. Zavodny, Madeline , “Welfare and the Locational Choices of New Immigrants,” Federal Reserve Bank of Dallas Economic Review, Second Quarter, 1997. Zavodny, Madeline , “Determinants of Recent Immigrants’ Locational Choices,” International Migration Review, Vol. 33, No. 4, 1999. 38 About the Author Sarah E. Bohn is a research fellow at the Public Policy Institute of California, where she studies immigrants, immigration policy, economic demography, and the economy. Her work focuses specifically on the effects of immigration on the labor market and on the economic progress of immigrants in the United States. She holds a Ph.D. in economics from the University of Maryland, College Park. PUBLIC POLICY INSTITUTE OF CALIFORNIA Board of Directors Walter B. Hewlett, Chair Director Center for Computer Assisted Research in the Humanities Mark Baldassare President and Chief Executive Officer Public Policy Institute of California Ruben Barrales President and Chief Executive Officer San Diego Regional Chamber of Commerce John E. Bryson Retired Chairman and CEO Edison International Gary K. Hart Former State Senator and Secretary of Education State of California Donna Lucas Chief Executive Officer Lucas Public Affairs Ki Suh Park Design and Managing Partner Gruen Associates Constance L. Rice Co -Director The Advancement Project Thomas C. Sutton Retired Chairman and Chief Executive Officer Pacific Life Insurance Company Raymond L. Watson Vice Chairman of the Board Emeritus The Irvine Company Carol Whiteside President Emeritus Great Valley Center PUBLIC POLICY INSTITUTE OF CALIFORNIA 500 Washington Street, Suite 600 San Francisco, California 94111 p hone: 415. 291.4400 f ax: 415. 291.4401 PPIC SACRAMENTO CENTER Senator Office Building 1121 L Street, Suite 801 Sacramento, California 95814 p hone: 916.440.1120 f ax: 916.440.1121 www.ppic.org" ["post_date_gmt"]=> string(19) "2017-05-20 09:39:29" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(6) "closed" ["post_password"]=> string(0) "" ["post_name"]=> string(8) "r_709sbr" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2017-05-20 02:39:29" ["post_modified_gmt"]=> string(19) "2017-05-20 09:39:29" ["post_content_filtered"]=> string(0) "" ["guid"]=> string(50) "http://148.62.4.17/wp-content/uploads/R_709SBR.pdf" ["menu_order"]=> int(0) ["post_mime_type"]=> string(15) "application/pdf" ["comment_count"]=> string(1) "0" ["filter"]=> string(3) "raw" ["status"]=> string(7) "inherit" ["attachment_authors"]=> bool(false) }