University of North Carolina at Chapel Hill
320 New East (CB# 3140)
Chapel Hill, NC 27599-3140
firstname.lastname@example.org or email@example.com
July 17th, 2014
This paper explores variation in the economic integration of immigrants across US metropolitan areas and tests a basic hypothesis that greater economic integration promotes regional resilience. Here we construct two quantitative indexes of occupational diversity as primary indicators of economic integration and develop a conceptual framework of social, economic and spatial factors that are likely to shape occupational diversity at the regional scale. We conduct an exploratory quantitative analysis in two steps. First, we model labor market diversity in 2000 with metro-level data drawn primarily from the Building Resilient Regions (BRR) database. Next, we use the occupational diversity indexes as dependent variables and assess whether greater occupational diversity among immigrants leads to greater economic resilience between 2000 and 2010, as measured by changes in unemployment rate and real wage growth. We find some evidence that immigrants in regions that have more broadly integrated immigrants (across occupations) were relatively more resilient in the face of the economic shocks of the Great Recession.
As the economy continues to recover slowly from the Great Recession, the flow of new immigrants into the United States has slowed significantly. Two decades earlier, millions of new workers and their families migrated to and settled in a variety of U.S. regions. This period of migration is markedly different in three key ways than other periods: 1) large scale immigration from Latin America and Asia – particularly Mexico, 2) a continual replenishment of immigrants from the same sending countries, and 3) migration to “new destinations” such as regions in the Southeast and new types of communities—suburbs and rural towns (Waters & Jiménez, 2005).
Thus, it is critical that policymakers understand how the U.S. economy performed in successfully integrating new migrants into the labor market and the degree to which these families are able to move up the economic ladder. Since the pattern of immigration was uneven across regions and also varied by country of origin and skill level, we might also expect that there is wide geographic variation in the level of economic integration of immigrants. Uneven patterns of economic integration may also be driven by factors relating to: 1) human capital; 2) the context of reception; 3) migration to ethnic enclaves; 4) spatial mismatch between where immigrants locate (e.g. concentrated poor neighborhoods in the central city) and where jobs are located (e.g. suburbs); and 5) the increasingly divergent patterns of economic development across regions, as some “innovative regions” move far ahead of declining metropolitans areas in wealth generation and economic opportunity (Moretti, 2012).
This paper will explore the patterns of immigrant economic integration across metropolitan areas with large immigrant populations in the U.S. and attempt to explain which factors or sets of factors are associated with immigrant economic integration. Furthermore, this study will explore whether immigrants in metropolitan areas with higher levels of economic integration are more resilient to the economic shock resulting from the Great Recession. Specifically, this paper will document the extent to which immigrants are relegated to isolated niches in the labor market that lack the opportunity for economic mobility—for example, in food processing occupations in the Southeast—or, conversely, well-integrated to a variety of occupations and industries throughout a regional economy. This paper builds upon research supported by the BRR network in the past; particularly in Pastor, Lester and Scoggins (2009), which demonstrates the divergent trend in regional performance, and Chapple and Lester (2010) which explores the factors behind resilient regional labor markets. In addition, this paper will use the Building Resilient Regions (BRR) database as a basis for explanatory variables and will add additional updated (e.g. 2010 American Community Survey data) measures to this shared resource. While previous studies have examined immigrant integration on a select number of comparison regions (Pastor and Mollenkopf, forthcoming) or regions within a single state (Pastor et al., 2012), this data allows us to examine a large number of metropolitan areas.
This paper is intended to be an exploratory analysis that highlights regional variations in the economic integration of new immigrants. We define and test several quantitative measures of occupational diversity among immigrants as a key proxy for their economic integration. Next, we explore the characteristics of regions that are associated with greater economic integration and test several leading theories (e.g. human capital, context of reception, ethnic enclave, spatial mismatch, and regional industrial structure). Finally, we test the relationship between economic integration of immigrants and regional economic resilience by measuring the effect of immigrant occupational diversity on unemployment and real wage growth before and after the Great Recession. The remainder of this paper is structured as follows: Section two reviews the literature on immigrant economic integration and develops a conceptual framework that describes the theoretical determinants of economic integration. Section three describes the methodology used to measure occupational diversity and outlines the quantitative analyses to follow. Section four presents the results and summarizes the findings. In section five we offer some discussion of our findings. The final section concludes with policy recommendations and outlines a research agenda that builds upon this exploratory research.
2. Immigrant Economic Integration in the U.S.
2.1 Classical and Segmented Assimilation
Classical models of immigrant integration or assimilation have heavily focused on European immigration to urban gateways, such as New York and Chicago, in the early 20th century. During this period, immigration occurred in large waves (as opposed to steady streams), which allowed for studies of immigrant cohorts. In general, studies revealed that there was a linear process of assimilation for European immigrants. Immigrants became more integrated with longer residence and were fully integrated into host societies after two or three generations (Alba and Nee 2004, Joppke and Moawska, 2003; and Ireland, 2004). This model predicts that higher levels of human capital, including English language ability, education, and work experience, will accelerate economic integration.
Post-1965 changes in Federal immigration policy that resulted in large-scale immigration from Latin America and Asia and changing settlement patterns challenged traditional assimilation theories. Scholars found that not all contemporary immigrant groups follow a linear assimilation process as posited by classical models, but rather, they follow divergent paths of assimilation (Rumbaut and Portes, 2001; Kasinitz et al., 2008). This alternative model, segmented assimilation, suggests that while some immigrants achieve socioeconomic mobility and assimilate into the middle-class, other groups experience “downward assimilation” leading to permanent poverty and spatial settlement with the underclass (Gans, 1992; Portes and Zhou, 1993; Zhou 1997; Portes et al., 2005), as in the case of West Indians in Miami, Florida.
Findings from empirical tests of these theories do not resolve the debate over the process of immigrant economic integration. On the one hand, a number of studies find continuing support for the classic assimilation model. For example, Kasinitz et al.’s (2008) study of second (including 1.5 generation) immigrants from a variety of different ethnic origins in New York finds that children of immigrants have more diverse occupational profiles than their parents (e.g. more integrated into the labor force). Furthermore, second generation groups that are more successful in the labor market, such as Chinese and Russian Jews, are more likely to assimilate into the mainstream economy. That is, labor force participation by second generation Chinese and Russian Jews most closely resembles native-born New Yorkers of their same race. For all second generation immigrants studied, their wages are more similar to native-born whites than native-borns of their same race (i.e. the comparison groups). Overall, Kasinitz and his colleagues concluded that downward assimilation was rare for second generation immigrants and that joining the mainstream economy, as opposed to the ethnic economy, was the most successful path towards economic mobility. Alba and Nee’s (2003) review of studies about second generation immigrants and occupational position also shows upward mobility for second generation immigrant groups. For example, Asian, European or Canadians, and South American children of immigrants have greater representation in executive, managerial, and profession jobs than third and later generation non-Hispanic Whites. Furthermore, immigrants who arrive in the US at a younger age become more fully incorporated into mainstream American society, including having higher rates of participation in white-collar occupations (Myers, Gao, and Emeka, 2009).
On the other hand, the path to upward economic and occupational mobility was not equally traversed by all immigrant groups. In particular, Mexican immigrants and darker skinned immigrants (e.g. Haitians and black Caribbeans) and their descendants have not fared as well. Haller, Portes, and Lynch’s (2011) analysis of the Children of Immigrants Longitudinal Study data shows evidence of a segmented assimilation pattern with some groups, such as Cubans-Americans achieving high occupational status, while Mexican-Americans experience downward assimilation. In addition, they found that Haitian immigrant’s higher educational attainment did not translate to higher occupational status while other groups, including Chinese- and Korean-Americans were able to realize occupation gains by being more educated. A number of other studies point to a duality in Mexican immigrants’ economic assimilation. When compared to other second generation immigrants, Mexican immigrants may not fare as well, but compared to third generation (or later) Whites or when compared to their own parents, they show great strides in rates of economic assimilation (Portes and Rumbaut, 2001; Alba and Nee, 2003).
The empirical results reveal that there is no single linear path of economic assimilation. They also reveal that race and ethnicity play a very important role in explaining rates of assimilation. Cuban-Americans in Miami as compared to Mexican-Americans in Los Angeles, for example, arrive in the US under different pressures and bring with them varying levels of human capital. Their environment, such as the neighborhoods where they locate and socioeconomic and political contexts of their local and regional area also influences how quickly they incorporate. We examine the role of ethnic enclaves and the context of reception for immigrant integration below.
2.2 Ethnic Enclaves and Context of Reception
The presence of an ethnic enclave where immigrants work or reside and how they are received by the host society has been shown to affect their economic integration. Wilson and Portes (1980) were the first to present an ethnic enclave model of economic assimilation that diverges from both the classical and segmented assimilation models. Studying Cubans in Miami, they argue that an alternative path to socioeconomic mobility exists. This path runs through employment in the ethnic enclave and relies on strong co-ethnic ties and group solidarity. This model suggests that immigrants do not have to assimilate into mainstream Anglo-American society or the mainstream labor market in order to succeed.
Ethnic enclaves are defined by the concentration of co-ethnics in space. These concentrated co-ethnic neighborhoods, which allow immigrants to preserve their culture, maintain community solidarity, and access social networks may be another avenue for achieving economic advancement and labor market integration. The case of Cubans in Miami is an example of how ethnic enclaves can buffer the transition resulting from migration and can provide kinship ties that insulate Cuban immigrants from downward assimilation (Portes et al., 2005).
The persistence of ethnic enclaves, the steady stream of new immigrants from the same sending communities (e.g. immigrant replenishment), and the high numbers of undocumented immigrants who are often transnational residents that want to eventually return to their home country, raises interesting questions about the need or the desire to integrate into middle-class white society in order to achieve economic advancement (Kasinitz, 2008). Bonacich (1973) describes the immigrant sojourner as someone who does not fully participate in the civic life of the host society because he does not consider it his permanent home. She describes middlemen minorities as occupying an intermediate role in the economy, such as someone in between the employer and the employee or the consumer and the producer, with the following characteristics: “…resistance to out-marriage, residential self-segregation, the establishment of language and cultural schools for their children, the maintenance of distinctive cultural traits (including, often, and distinctive religion), and a tendency to avoid involvement in local politics except in the affairs that directly affect their group” (Bonacich, 1973, p. 586). Thus, sojourners and middle-men minorities are able to succeed economically, but do not depend on integration into the host society to do so.
While some racial/ethnic groups do not want to integrate, others face barriers in doing so, such as Black immigrants (Freeman, 2002). Studies suggest that living in racially or economically homogeneous neighborhoods can inhibit socioeconomic mobility, by restricting an individual’s social network to those who have similar resources and skills. Granovetter (1973) explains that it is not these strong ties with one’s interpersonal network in homogeneous neighborhoods that lead to employment opportunity, but weak ties (i.e. with acquaintances) that expand an individual’s connections to a more varied set of institutions and organizations. Thus, close-knit networks, such as those found in ethnically homogeneous neighborhoods (e.g. ethnic enclaves), that have strong “bonding capital” but little “bridging capital” can inhibit economic integration (Granovetter, 1973; Lin, 2000; Putnum, 2001). In addition, Hendricks (2002) finds that employers use race or ethnicity as a predictor of skill since it is difficult to evaluate new immigrants’ skills. This may account for why Haitians do not translate higher education into better occupational attainment and why Mexican immigrants do not achieve economic assimilation as rapidly as other immigrant groups (Portes and Rumbaut, 2001; Alba and Nee, 2003; Haller, Portes, and Lynch, 2011). Put differently, racialization in the labor market may exist for certain immigrant groups. Furthermore, the continual streams of immigration from poor sending countries can depress wages for immigrants entering in earlier periods and may also discourage earlier immigrants to invest in skills improvement because employers hire and pay on the basis of ethnicity and not necessarily skill level.
For some groups, solidarity with co-ethnics and ties with the ethnic enclave may result in economic advancement, but how immigrants are received by the host society may also play a factor. A burgeoning literature suggests that street-level bureaucrats, local immigration policies, immigrant advocacy organizations, and local government actions can work to facilitate or discourage immigrant integration (Jones-Correa, 2008a; Jones-Correa, 2008b, Marrow, 2009; Marrow, 2011; Steephen et al., 2013; Nguyen et al., forthcoming]. Thus, it is important to understand the context of reception across different localities.
2.3 Spatial Mismatch
While much of the theoretical debates over immigrant integration have been aspatial, there is a growing body of research that applies Kain’s (1968) spatial mismatch thesis to immigrants. Kain’s seminal study on housing segregation, decentralization of jobs, and Black employment found that Blacks living in concentrated poor neighborhoods in the central city were disconnected from major growth centers (e.g. suburbs). Thus, residential segregation of Blacks in urban areas and job growth in the suburbs, otherwise known as the jobs/housing imbalance, results in higher overall unemployment and greater poverty for the region. While the magnitude of the effect of employment decentralization on Black unemployment has been debated, decades of research offer evidence that Blacks in the central city have less access to jobs than Blacks and Whites in the suburbs and Blacks who are employed have higher commute times than employed Whites (Holzer 1991).
Recent migration trends show that immigrants are bypassing immigrant gateways and locating in new destinations and even locating directly to suburbs rather than central cities (Frey, 2003; Singer, 2004). This raises the question of whether job decentralization has impacted immigrants differently than Blacks. Lui and Painter’s (2012) study of sixty metropolitan areas finds that immigrants are more spatially segregated from jobs than Whites, but less so than Blacks. Furthermore, they find that immigrants are more residentially mobile than Blacks, and can thus follow the jobs, while Blacks are slower to locate residentially to where the jobs move. Another study also found that first generation Latino immigrant youths’ employment was not constrained by whether they lived in the central city, inner-ring suburbs, or outer-ring suburbs, suggesting that Latino youth are more residentially mobile than even White youth (Painter et al., 2007).
2.4 Occupational Diversity of Immigrants and Resilience
Significant attention has been paid to defining the concept of “resilience” and understanding how to operationalize it (Christopherson et al., 2010; Foster, Pendall, and Cowell, 2010; Simmie and Martin, 2010). In this paper, we define resilience as the ability of metropolitan areas to be more or less adaptable to economic stress caused by the Great Recession. We hypothesize that immigrants in regions with greater occupational diversity of immigrants will be more resistant to and recover more quickly from the economic downturn. Specifically, we examine whether regions that have higher levels of occupational diversity are more resilient to the economic shocks posed by the Great Recession. In particular, we test the relationship between occupational diversity among immigrants in 2000 and the resulting change in unemployment and real wage growth between 2000 and 2010 among immigrants.
3.1 Measuring Occupational Diversity
There are a number of ways to conceptualize and operationalize immigrant economic integration. Economic integration could mean achieving a middle-class standard of income and be measured in a simple uni-dimensional way. However, income alone does not fully capture the concept of “integration.” Alternatively, economic integration could be measured in a broad, multidimensional manner by using a combination of indicators, such as income, home ownership, and children’s educational attainment. Furthermore, these indicators could be measured over time for individuals or across different generations, thereby requiring longitudinal measures. However, longitudinal data is extremely difficult to obtain, especially at the regional level. Thus, we chose a measure of economic integration that was richer than simply measuring income but also could be measured for all regions in the U.S. Specifically, we use labor market diversity as our primary proxy measure for economic integration of immigrants. This measure is supported by Waldinger (2005) who argues that newly arrived immigrants tend to cluster in occupational niches, but when they make economic progress and assimilate, they disperse throughout the labor market.
Labor market diversity is referred to here as the degree to which immigrants are employed throughout a region’s economy and are not simply concentrated in a few key sectors or niches. We believe this adequately captures the concept of economic integration, even if it only captures one aspect of it. Consider a hypothetical region where recent immigrants are recruited for agricultural work and are not able or allowed to work in other sectors, even those where their skills could be relevant. Immigrants in this region would lack economic integration under our conceptualization in that they are less likely to find out about opportunities to advance in other sectors or interact with non-immigrant workers (to gain greater knowledge of the labor market and U.S. institutions). Compare this region to another where immigrants are able to find employment in a broad spectrum of occupations. In this region, a worker is more likely to find a good match with his or her skills. Thus, while labor market diversity is not the perfect or the only measure of economic integration, we argue that it captures some critical elements.
To measure the economic integration of immigrants we constructed two distinct measures of occupational diversity at the metropolitan level. First we construct a non-relative occupational diversity index based on the Hirshman-Herfandahl Index.
As described in equation 1, the occupational diversity index is defined for each metropolitan area (i) and is based on the squared shares of workers in each occupational category () compared to the overall workforce. We defined nine broad occupational categories based on the Integrated Public Use Microdata Sample (IPUMS1) variable ‘occ’. The occupational diversity index is calculated separately for three groups (j): all immigrants, Mexican immigrants, and native-born workers. Thus, the term for the occupational diversity index of immigrants represents the share of immigrants in occupational category k out of the total number of immigrant workers in the metro area. If all immigrant workers were concentrated in only one category, then occdivi would equal zero (i.e. 1 – 12=0). Alternatively, if workers were evenly distributed across all categories the index would equal 1- (1/k), or 0.889. Thus higher values of the diversity index indicate more diversity across occupations, while lower values reflect more concentration.
As with all categorical measures of diversity, our occupational diversity index is highly dependent on the number of categories and the method used to develop them. There is an inherent tension in developing the occupational categories between the level of detail achieved and the statistical limits of the microdata samples we employed. On the one hand, we would ideally like to capture the degree of immigrant concentration in key occupations that are dominated by immigrants, at least anecdotally (e.g. restaurant cooks, drywall installers, etc.) However, if we use too many occupations we will not have sufficient sample size in each metro/occupational cell to estimate an accurate measure of . Thus, our goal in determine a categorizing scheme was to maximize the sample size of regions for which we can accurately calculate the indices, while also capturing important aspects of labor market segmentation. Ultimately, we used the following categorization scheme, which roughly approximates the major occupational groups defined in the Standard Occupational Classification (SOC) system: 1) management; 2) professional, technical and protective services; 3) low-wage services (includes food services, home health aids, building maintenance occupations); 4) sales and office/administrative; 5) agriculture; 6) construction; 7) other blue-collar jobs (includes transportation, utilities, communication, repair, and resource extraction occupations); 8) production and/or manufacturing; and 9) military and unclassified occupations. We made minor modifications that better approximate the skill and wage distinctions within the service sector. Table 1 in the next section lists the distribution of employed workers across these occupational categories for immigrants, native-born workers and Mexican immigrants. It is important to note that we do not include self-employed individuals as a separate category. While the literature suggests that access to entrepreneurship is an important indicator of economic success, sample size limitations at the metropolitan level preclude this analysis.
This way of measuring diversity does not make comparisons in a given metropolitan area to a reference region (e.g. the U.S. as a whole). As such, it simply measures diversity across a given set of categories within a single economy. There is no implied “ideal” distribution across occupations since the structure of labor demand is itself likely to vary across metropolitan areas for reasons that do not relate to the degree of integration of immigrants. Thus, one concern in conceptualizing the occupational diversity index as a measure of immigrant economic integration is that the index may simply be reflecting variation in the industrial structure across different metropolitan areas. For example, large metropolitan areas with diverse economies naturally have a greater opportunity for immigrants to be distributed across the broader set of occupations since they have industries that contain these occupations. Conversely, smaller regions that are economically specialized in one major industry may simply not have the labor demand for workers (immigrants or otherwise) in certain occupations. While, we recognize that this is a concern, in the main empirical analysis below we explicitly control for the industrial diversity of each metropolitan area when analyzing the impact of diversity on key outcome variables. Second, since this index is calculated for occupations, rather than industries, and is measured at a relatively coarse number of categories, it is still effective at capturing the relative level of opportunity that immigrants are granted within the regional labor markets. Furthermore, even regions that are relatively specialized still have some amount of employment in low-wage, residentiary industries (e.g. restaurants) and construction, and the occupational diversity index would capture the degree to which immigrants are overly concentrated in these occupational groups or not.
for differences in industrial structure across regions, we calculate an alternative measure of occupational diversity.2 Specifically, we use an index of specialization. As equation 2 indicates, the index of specialization is a relative index that compares the share of immigrant workers employed in a given category k in each region i to the same share for all workers of group j (e.g. immigrants, native-born workers, etc.) across the US as a whole.
The index of specialization (ISij)in equation 2 ranges from zero to one, with more diverse regions closer to zero and more specialized ones closer to one. A higher index of specialization means that the absolute differences between the share of workers in the occupational groups is higher. In other words, higher values indicate that the immigrant labor-force in the region is more specialized or concentrated in some categories compared to the distribution of immigrant workers overall.
3.2 Data Sources and Construction Steps
The primary data source for the measures of occupational diversity are the Integrated Public Use Microdata Sample (IPUMS) files from the U.S. Census Bureau maintained by the University of Minnesota Population Center (Ruggles et. al., 2013). We used microdata extracts from the 5 percent (long-form) 2000 decennial census and the 2010 American Community Survey (ACS). In addition to the occupational diversity indices, we also computed the percent of immigrant workers with a Bachelor’s degree or higher, the rate of unemployment and the real income growth of each immigrant group for each year. All values were calculated at the metropolitan area level using the consistent IPUMS variable metarea and then rescaled to the current combined statistical area (CBSA) definitions.3 We then merged these variables with a selected subset of relevant variables from the Building Resilient Regions (BRR) database (Pastor, Lester, and Scoggins, 2009). The BRR database was developed by a MacArthur Foundation funded research network and contains over 1,400 variables that measure a wide variety of demographic, economic, social, and political characteristics of metropolitan regions. While the database contains information for all metropolitan areas in the U.S., some variables derived from microdata were only available for larger regions (i.e. those with a minimum of 200,000 persons in 2000). Thus, we limited our analysis to the sample of 192 metropolitan areas that meet this size criterion.
We also developed a proxy measure to test for the ethnic enclave effect. Specifically, we used tract level data from 2000 on nativity status and country of origin to calculate the percent of the foreign-born in each tract. Tracts that had at least 30 percent immigrants (or Mexican immigrants) were considered enclaves. We then calculated the proportion of a region’s immigrant population that lives in an enclave. This captures the degree to which immigrants living in a region are spread out throughout the metropolitan area, or living in close proximity to one another.
In addition to using data from the IPUMS and BRR databases, we developed two new variables which we argue are metrics of a region’s context of reception of immigrants. These variables are policy or civic variables that are intended to be proxies for the broader institutional setting at the regional scale. In the last decade, a greater number of local elected officials and local governments have attempted to actively facilitate immigrant integration by changing public and administrative policies and institutionalizing the change through the development of immigrant service offices. For this study, we identify the metropolitan areas that have an immigrant or ethnic service office or have adopted immigrant-friendly city-wide initiatives as “pro-immigrant.” We classify metropolitan areas as “anti-immigrant” if they have adopted policies that immigrant-advocacy groups have considered hostile to immigrants, such as the ICE ACCESS 287(g) Program. Both “pro-immigrant” and “anti-immigrant” are dichotomous variables and a single metropolitan area might have both types of policies.
To identify these policies, we searched government websites for the primary city or cities of each metropolitan area. We searched on the following keywords/terms: immigration, immigrant, and citizenship. We also looked at all the local commissions and the various types of mayor's offices (eg. Mayor’s Office of Immigrant Affairs, Mayor’s Office for New Americans) to see if they explicitly served immigrants, through programs such as translation for Spanish language speakers or an advisory commission for a specific ethnic group. We also examined special local initiatives, such as welcoming immigrant initiatives.
It should be noted that the quality and ease of navigating the city websites varies widely, so we may have missed a program/policy that was not easy to find on the website. On the flip side, this may be indication of the importance of the initiative. If it is difficult to find publicly, there may not be much public support for the initiative. These are somewhat crude measures of pro- and anti-immigrant policies, but there is currently no centralized database that contains information on the context of local government reception for immigrants.
Lastly, we used data from the National Center for Charitable Statistics (NCCS) database from the Urban Institute to count the number and revenue of non-profit organizations that serve immigrant or ethnic populations. These immigrant-serving organizations may provide a broad array of services to immigrant communities including housing and social services in addition to labor market support. These variables were normalized by the total immigrant population in the given year.
3.3 Research questions and empirical strategy
The main empirical analysis in this paper consists of two distinct tasks. First, we attempt to explain the variation in occupational diversity across metropolitan areas based on the conceptual framework developed in section two. In this task we ask which factors are associated with occupational diversity, our main proxy for economic integration of immigrants. For each of the potential explanatory factors—human capital, ethnic enclaves, spatial mismatch, context of reception, and regional industrial structure—we chose a set of metrics from our updated BRR database (see Appendix A for the list of specific variables chosen to proxy for each factor). Next, we estimated a set of OLS regression models that measure the association between each factor and our various measures of occupational diversity (the dependent variable). It is important to note that our purpose in conducting this OLS analysis is not to test specific causal relationships, but rather to explore broad associations between regional characteristics and occupational diversity.
The next task is to test the hypothesis that occupational diversity among immigrants leads to greater resilience among immigrants at the regional scale. For the purposes of this paper we define resilience in a strictly economic sense—the ability to withstand economic shocks such as the Great Recession—and use two primary outcome measures of resilience: the change in the unemployment rate of immigrants and the change in real wage income of immigrants between 2000 and 2010.