TY - JOUR
AU - Zhou,, Bo
AB - Abstract Building on the growing literature on new immigrant destinations, this paper examines new employment patterns of low-skilled Chinese immigrants in the United States. We identify an important channel of employment in new destinations for the case of Chinese low-skilled immigrants: employment agencies in New York City’s Chinatown. We carried out two surveys of employment agencies during 2010–2011. Our findings suggest that there has been a profound change in settlement patterns of low-skilled immigrants: moving away from traditional Chinatowns in major American cities toward non-gateway destinations and rural areas. These new settlement locations are characterized by a low unemployment rate and low crime rate. Contrary to predictions from ethnic economy and mainstream economic perspectives, Chinese restaurant jobs tend not to be in places with a high concentration of Chinese immigrants, but rather in places with a high proportion of non-Hispanic whites. In addition, the farther the jobs are from New York City, the higher the salary. We discuss the implications of this fundamental change for re-conceptualizing the immigrant labor market and immigrant socioeconomic mobility in American society. Introduction On January 28, 2005, the Wall Street Journal ran a front-page article titled “On the East Coast, Chinese Buses Give Greyhound a Run” (Newman 2005). It focused on how bus companies operated by Chinese immigrants compete with Greyhound buses for customers on the East Coast routes. The story took many by surprise because it seemed inconceivable that immigrant-owned businesses could compete with big-name corporations like Greyhound (worth $1 billion at that time). However, underneath this sensational report lies a sociological story of recent Chinese immigration and settlement to non-gateway destinations in the United States. Located in Manhattan’s Chinatown, the Chinese bus companies, also known as Chinatown buses, were created to serve the needs of recent Chinese immigrants who found jobs in Chinese restaurants. These jobs were located far away, typically in non-gateway destinations such as Rhode Island, Maine, or Alabama. According to the Wall Street Journal article, these companies gained solid footing in the business and expanded their customer base beyond Chinese immigrants due to their flexible schedules, convenient pickup/dropoff points, and less expensive bus fares (also see a more recent report from the New Yorker [Hilgers 2014]). Immigration researchers have long taken for granted that immigrants concentrate in major gateway cities and states. Historically, for many Chinese immigrants, going to the United States was almost synonymous with going to the Gold Mountain (San Francisco). Traditional settlement cities in many ways symbolize the meaning of America. Well-cited seminal studies tacitly embraced and perpetuated this attribution by focusing on major metropolitan areas with large numbers of immigrants (Foner 2000; Logan, Alba, and McNulty 1994; Logan, Alba, and Zhang 2002; Nee and Nee 1973; Nee, Sanders, and Sernau 1994; Portes and Bach 1985; Waters and Jimenez 2005; Zhou 1992). Some even observed that post-1965 immigrants concentrated even more in selected locations than previous immigrant cohorts (Massey 1995; Min 2005). Recently, researchers noticed an emerging phenomenon of immigrants arriving at non-traditional destinations. Using data from the US Census, Singer (2004) showed that nearly one-third of immigrants resided outside established settlement states in 2000, and subsequent studies have produced significant insights into the settlement process in new destinations (Gozdziak and Martin 2005; Hirschman and Massey 2008; Marrow 2011; Massey 2008; Zuniga and Hernandez-Leon 2005). However, research has focused almost exclusively on Mexican and Latino immigrants (though see Flippen and Kim [2015], who study Asian Americans, as a notable exception). We argue that it is equally important to carefully examine other groups, who may reveal different patterns, so that we may consolidate, or perhaps challenge, our understanding of the new settlement process. In this paper, we study low-skilled Chinese immigrants who have recently arrived in new immigrant destinations. Researchers underscore two major justifications for studying the settlement of immigrants in new destinations and their future for assimilation. First, new destinations often lack the institutions usually associated with successful assimilation, such as ethnic-based churches, bilingual education for children, bilingual services (i.e., hospital, Department of Motor Vehicles, and Emergency Medical Services), sizable ethnic communities (i.e., Chinatowns or Koreatowns), and immigrant-based organizations (hometown associations). In traditional destinations, these institutions can ease the initial settlement process and ultimately help immigrants assimilate into American society. Second, researchers also note that residents in non-gateway destinations usually do not have prior encounters with immigrants, which could lead to potentially troubled race and ethnic relations (Massey 2008; Waters and Jimenez 2005). While acknowledging that these arguments likely hold for Chinese immigrants, we argue that the study of Chinese immigrants in new destinations brings additional theoretical and empirical significance. A major trend that characterizes recent settlement of Chinese immigrants in non-gateway destinations is that the process involves both Chinese employers (restaurant owners) and employees. Business ownership itself can be a measure of success, and the expansion of Chinese restaurants to non-gateway destinations provides a new (perhaps faster) avenue for economic mobility. This expansion also challenges the consensus that ethnic economies need to remain concentrated (Logan, Alba, and McNulty 1994; Portes 1995). How Chinese immigrant owners operate and adapt in such a new environment is not well understood. Furthermore, the recruitment mechanism for moving to new destinations may differ across ethnic groups. For the Chinese case, there is heavy reliance on employment agencies (EAs) operated by Chinese immigrants and located within New York City’s (NYC) Chinatown. The EAs serve as a link by providing critical information about the new destination and employers to potential employees. Perhaps more than realized, EAs facilitate the settlement process for employees who venture into the new destinations. Thus, a systematic examination of new patterns of employment and settlement among recently arrived low-skilled Chinese immigrants is clearly needed. Using NYC as a research site, we join research efforts to examine patterns of settlement and assimilation among recent Chinese immigrants in non-gateway settings. We have three objectives. First, we introduce two major actors that facilitate the process of Chinese immigrant settlement: employment agencies and Chinatown buses. We realize that the emergence of these two facilitators is a profound development within the Chinese immigrant community, deserving a paper-length analysis in its own right. In this study, we present an abbreviated treatment of the emergence of these drivers of immigrant non-gateway settlement. Second, we consider contextual-level factors that are conducive for Chinese immigrant entrepreneurs’ business expansion into non-traditional destinations. Third, we explore consequences of the new settlement pattern for immigrant adaptation and economic mobility by focusing on linkages between immigrant wages and geographical location. We end the paper with implications for re-conceptualizing immigrant labor market and immigrant socioeconomic mobility in American society. Background and Hypotheses Review of Previous Studies Historically, immigrants to the United States settled in several urban destinations. This is partially due to the role of migration networks, which channeled new immigrants to places where earlier cohorts settled. Immigrant settlement largely occurred in urban areas, because of employment opportunities and a more receptive nature for diversity. However, the 1990s signified a major shift in immigrant settlement patterns: a large number of immigrants moved to new destination cities such as Atlanta, Las Vegas, and Denver, and a significant volume of immigrants moved to non-metropolitan areas as well (Donato et al. 2008; Singer 2004).1 Using data for recently arrived immigrants from censuses and the American Community Survey (in the five years prior to the census date), Massey and Capoferro (2008) provide perhaps the most systematic portrait of the geographic diversification of American immigration. In 1990, 35 percent of recent immigrants resided in California; however, by 2000, that percentage had declined to 23 percent. Likewise, 13.4 percent of immigrants resided in New York in 1990, but it dropped to 5.9 percent by 2000. Among all immigrant groups, perhaps the most dramatic drop is for Mexicans, whose percentage in California dropped from 63 percent in 1990 to 33 percent in 2005. This dramatic shift in the settlement patterns has stimulated increasing research. Massey and Capoferro (2008) suggested four explanations for the diversification of settlement patterns. The first factor focuses on the Legalization Program from the Immigration Reform and Control Act (IRCA) of 1986, which resulted in the saturation of the labor market, especially in California. Second, the passage of Proposition 187 made California a less welcoming environment for immigrants. The third factor was a “selective hardening of the border” that deflected immigrants to other destinations. The last factor was the changing geography of labor demand resulting from production restructuring. The restructuring of production often means deunionization of the workforce, subcontracting of labor, and relocation of plants to non-metropolitan areas to avoid unions. As a result, jobs become less attractive to native workers and immigrants become a more reliable and flexible workforce. Several recent studies provide evidence consistent with this perception (Leach and Bean 2008; Parrado and Kandel 2008). In the case of California, Light’s (2006) study points to the role of local governments in creating an unwelcome context in metropolitan areas, which encourages relocation of immigrants to non-gateway destinations. In sum, previous studies have clearly documented patterns of geographic diversification of recent immigrants, advanced several explanations, and presented initial evidence for this diversification. Researchers have also expressed concerns for the prospects of immigrant assimilation in new destinations (Fennelly 2005; Lichter et al. 2010; Marrow 2011; Waters and Jimenez 2005). We note that much of this work has centered on the experiences of Mexican and Latin American immigrants, with one major exception, Flippen and Kim (2015). In 2007, 41 percent and 34 percent of immigrants came from Latin American and Asian countries, respectively (Office of Immigration Statistics 2008). By 2015, the number for Asian immigrants had risen to 39 percent (Office of Immigration Statistics 2016). Thus, it is important to see if there is a geographic diversification story for Asian immigrants. Indeed, Flippen and Kim (2015) find that moving to non-traditional destinations has different consequences within Asian groups. For example, for Chinese immigrants, residence outside traditional destinations is associated with more favorable socioeconomic outcomes; however, this is not the case for Koreans and Vietnamese. Building on this body of literature, this paper examines the geographic diversification of recent Chinese immigrants using innovative data collected in NYC’s Chinatown. Chinese Immigrants in the NYC Context For a long time, researchers needed only to go to Chinatowns in different cities to study Chinese immigrants. However, this strategy no longer accurately captures contemporary Chinese immigrants in the United States. For recently arrived low-skilled Chinese immigrants, most of the employment opportunities are located outside Chinatown. This paper provides a more nuanced analysis of the settlement process in non-gateway destinations. We argue that this process results from the saturation of the Chinese restaurant market, which was a job machine for generations of Chinese immigrants in NYC, and a corresponding increase in job demand due to a rise in immigration from China’s Fujian province (Liang et al. 2008). By the time many Chinese immigrants arrived in NYC in the 1990s, there were no longer places to open new Chinese restaurants, with some estimating that there were as many as 2,000 Chinese restaurants in the NYC region by the early 2000s. In addition, commercial rent rapidly increased, thus lowering the potential profit margin and further dimming the prospect of opening a restaurant. As such, Chinese immigrant entrepreneurs face a competitive market if interested in opening their own restaurant. Determinants of Restaurant Job Locations: Previous Studies Historically, ethnic businesses often began by serving the needs of co-ethnic groups in the form of either restaurants or grocery stores (Portes 1995). In this regard, the close proximity of ethnic businesses to ethnic neighborhoods was important. Ivan Light and colleagues carried out a series of studies of immigrant entrepreneurship among Korean, Chinese, Japanese, and Iranian immigrants (Light and Bonacich 1988; Light and Rosenstein 2005; Light et al. 1994). In most cases, their studies reveal that spatial clustering of immigrant businesses (often located in immigrant enclaves) is a key characteristic of these businesses. For example, even when Korean immigrants in Los Angeles move their businesses to the suburbs, their businesses are still spatially clustered and continue to rely on Korean immigrant workers (Light and Bonacich 1988). Light et al. (1994) identified an exception: Iranian immigrant business owners in Los Angeles. Iranian immigrant business owners are not spatially clustered, in part because most of them are self-employed and do not need to employ other Iranian immigrant workers. Recent studies of ethnic economies provide evidence that there are advantages in utilizing the ethnic social capital embedded in immigrant neighborhoods because social capital fosters and facilitates entrepreneurship. One of the challenges for immigrants starting their own businesses is accessing financial capital. Many immigrants lack the good credit necessary to obtain loans from mainstream financial institutions, so they rely on co-ethnic members to pool financial resources to start their businesses. The credit rotation practice has long been noted and continues today (Granovetter 1995; Portes and Sensenbrenner 1993).2 Perhaps the chief advantage of locating businesses near immigrant neighborhoods is easy access to immigrant labor (Fong, Luk, and Ooka 2005; Sassen 1995; Zhou 1992). Fong, Luk, and Ooka (2005) suggest that the human ecology perspective predicts that immigrant businesses are likely located in a city’s “transition zone,” which is characterized by a high level of poverty and social disorganization. However, findings from their study in Toronto are not consistent with this perspective. Another perspective draws insight from economic geography. Rooted in the new industrial space thesis (Scott 1988), Zhou (1998) studies the role of inter-firm networks in the location strategies of Chinese producer service firms in Los Angeles. For example, Chinese-run computer businesses are typically located at the fringe of Chinese-concentrated areas while maintaining proximity with other computer distributors so that parts can be exchanged faster and more efficiently. Both Fong, Luk, and Ooka (2005) and Zhou (1998) note this shift in Chinese business patterns of settlement from ethnic enclaves, such as Chinatown, toward the city outskirts or suburban areas, clearly signaling a departure from the traditional business settlement patterns. This pattern can be explained in part by the increase of Chinese immigrants living in suburban areas or ethnoburbs (Li 1998; Zhou 1998). There is also a fundamentally demand-side explanation advocated by economists that stresses the economic indicators of destination locations: unemployment rate, income level, and business climate (Card and Lewis 2005). This line of reasoning posits that entrepreneurs are more likely to open new businesses in locations undergoing economic growth and rising employment. This perspective is consistent with some of our interviews with restaurant owners, as they tend to identify “good locations” (hao qu) for their business operation. Employment Agency, Recruitment of Immigrant Workers, and Chinatown Buses Current perspectives on business location choices are not entirely satisfactory in explaining the case of Chinese restaurants. We extend our inquiry on location of businesses by including the settlement of Chinese restaurants in remote places. The recruitment of workers is a critical function for any business and may be more so for ethnic businesses. This is typically not an issue when the business is in a traditional gateway destination, because an abundant supply of immigrant workers is available. However, there is a spatial disconnect between Chinese businesses in faraway places and their supply of immigrant workers in gateway locations (i.e., NYC). The emergence of EAs in Manhattan’s Chinatown fills this gap. The current literature on non-gateway destinations rarely discusses the recruitment process for immigrants.3 In the case of Chinese immigrants, EAs fulfill this function. For example, in the three or four blocks around East Broadway and Eldridge Street in Manhattan’s Chinatown, there are 30–35 EAs. These EAs are important facilitators for the settlement process in new destinations. EAs resemble labor market intermediaries or brokerages in the mainstream economy (Autor 2008; Stovel and Shaw 2012). As Stovel and Shaw (2012) indicate, “successful brokerage results in the flow of some goods, service, or information from one party to another” (147). The key benefit of EAs is that they save both employers and employees time that would have been spent either searching for workers or searching for work (Autor 2018; Williamson 1975). Particularly in the case of workers, they no longer need to look through job websites or newspapers every day. In addition to serving the needs of Chinese immigrant workers and immigrant entrepreneurs, the emergence of EAs also responds to the increasing diversity of Chinese immigrant origins. Province origins are no longer restricted to Guangdong, which historically was the most important immigrant-sending province, but now include provinces such as Fujian, Zhejiang, and the northeast regions of China. Thus, traditional kinship and village-based migration networks are limited in this diverse environment and market-based job search agencies can cast a much bigger net. In addition, as the number of EAs increase in Chinese communities in NYC, their inflation-adjusted fees have reduced significantly. Mr. Yang, an EA owner in Chinatown, stated that when he came to the United States in 1973, the fees were $30 per transaction. Today, EAs still charge about $30–$35 per job transaction. As such, this price has remained nearly unchanged for more than 40 years! The current fee structure makes it easier for immigrant workers to look for or change jobs. More importantly, staff members in EAs provide needed information about non-gateway destinations and often work with bus companies to help immigrants learn how to travel to new restaurants or other job opportunities. Since they have the most up-to-date information on the restaurant job market, EA staff can also relay information about new locations and suggest routes to bus companies. Although EAs in Chinatown provide information to potential immigrant workers, there is still the challenge of spatial mismatch. Most of the newly established Chinese restaurants are in remote places far from Chinatown and the potential immigrant workers concentrated in the NYC area. The mismatch between jobs and workers has been a central focus of recent studies on minority groups. This mechanism was proposed as one of the leading causes of unemployment for some minority and immigrant groups (Kasinitz and Rosenberg 1996; Mouw 2002; Wilson 1987). For example, Wilson (1987) argues that the unemployment rate for blacks is high because of the spatial mismatch between inner-city blacks and job growth in suburban areas. In the case of Chinese restaurant jobs, Chinatown buses were created to solve the spatial mismatch of jobs and employees by providing convenient transportation for restaurant workers. As such, Chinese immigrant entrepreneurs help maintain job growth by opening new restaurants and providing a transportation network to connect recent immigrants with those job opportunities (New York City Department of City Planning 2009). As Chinatown buses continue to expand, many other minority groups, especially African Americans and Latinos, have become frequent passengers. Perhaps the most sociologically interesting function of EAs is their insertion of market mechanisms into the settlement of immigrants. The traditional way of employment/settlement is more kinship/relative-based (i.e., immigrants work for relatives or relatives’ friends, and thus may be exploited but cannot complain or relocate; Sanders and Nee [1987]).4 The introduction of EAs has fundamentally changed this reliance on relatives for immigrants by providing more choices in terms of job locations, types, and salary scales because all job-related information is posted on the bulletin boards of EAs. Jobs in non-gateway destinations may look especially attractive because most employers provide room and board, whereas jobs in NYC restaurants provide only free food. However, our recent interviews with Chinese immigrant workers reveal that they prefer to work in NYC to be close to their families/relatives, friends, and immigrant community. As such, immigrant salaries in non-gateway destinations closely reflect market prices. This means that if there is a short supply of workers in new destinations, the wages are set higher to attract immigrant workers from the allure of NYC. With respect to linking wages with immigrant enclave employment, a heated debate emerged in the 1980s and 1990s about the economic benefits of working in immigrant enclaves. Portes and colleagues (Portes and Shafer 2004; Wilson and Portes 1980; Zhou 1992) argue that immigrant workers in the enclave receive higher returns on education than those working outside the enclave. In contrast, Sanders and Nee (1987) show that the enclave hypothesis is only true for employers. A longtime Chinatown observer, Peter Kwong (1997), goes further to suggest that Chinese immigrant workers are exploited by their Chinese immigrant business owners. We add a new spatial dimension to this old debate by looking at immigrant workers who work in a gateway city (NYC) versus those in faraway locations. In both cases, the actors are still Chinese immigrants, but the job locations have changed. Chinese restaurants’ chefs provide a good example of how faraway locations shed new light on the economic benefits debate. A recent report suggests that chefs in non-gateway locations are paid about $2,300–$2,700/month while chefs in NYC are only paid about $2,000–$2,300/month despite the higher cost of living (Zeng 2009). Employers who offer below-market wages are less likely to receive job applications when the salary level of other job listings is known. Thus, we expect an association between the spatial distance of jobs and the salary offered, with the farthest jobs offering the highest wages. Furthermore, job mobility is increasingly possible because immigrants can leave their current jobs for other higher-paying jobs without constraints imposed by kinship/relatives. Most employers are close to their employees and even provide shuttle service for going to work. This closeness and interaction may enhance the opportunity for employees to learn business tips from their bosses, and eventually start their own business. Thus, the training system for entrepreneurship may work more effectively in new destinations than in NYC, where an employer’s main concern is that workers get their jobs done (Bailey and Waldinger 1991). In this paper, due to data constraints, we only test the wage differentials between immigrants working in NYC versus immigrants working in non-traditional destinations, and we await the testing of other ideas in future research. Data and Methods Our paper draws mainly on original data we collected about employment agencies in NYC’s Chinatown. We supplement this data when needed with existing data. We carried out in-depth interviews with staff members/owners of EAs, owners of selected Chinatown bus companies, and immigrant workers searching for jobs through EAs. Some of the information gathered from our interviews is used in this paper. Statistical analysis and models aim to examine key contextual-level factors that determine restaurant job locations and the extent to which immigrants who work in non-traditional destinations have a more favorable outcome in salary. We take advantage of the spatially explicit data and employ spatial modeling techniques in addition to statistical modeling. Data Sources (1) Job Listings from EAs In September 2010, we surveyed EAs in Manhattan’s Chinatown. The geographic locations of EAs are shown in map 1. For certain locations, one data point represents more than one EA because some buildings have multiple EAs within them. EAs are mainly located on Eldridge Street, Division Street, and East Broadway on the Lower East Side of Manhattan. Some informants call East Broadway “Fuzhou Street” to represent the large number of immigrants from Fujian province there (Liang et al. 2008). The geographic concentration of these EAs is not random but emerged because they initially wanted to serve the employment needs of Fujianese immigrants. Today, EAs also serve employment needs for Latino workers (Dolnick 2011). In September 2010, we surveyed 11 EAs out of the 32 located in Manhattan’s Chinatown. For each EA, we recorded job-related information, including location (determined by telephone area code), salary level, type (chef or food delivery), work hours, type of restaurant (most were Chinese, but a few served Japanese food), and any other pertinent information (e.g., how tips are distributed, any preference for immigrants who came from different parts of China). We obtained permission from each EA’s staff members to collect this information. During quiet business hours, graduate students visited each of these EAs and obtained information on each job posted at that time. We gathered information on 2,147 jobs. To capture potential seasonal variations on the job market, we repeated this process in February 2011 with a new random sample of 11 EAs. Information on 2,316 jobs was gathered during the second survey. Map 1 View largeDownload slide Location distribution of employment agencies in Manhattan’s Chinatown Map 1 View largeDownload slide Location distribution of employment agencies in Manhattan’s Chinatown (2) Area Code Data and County-Level Data We use area-code level as our unit when analyzing business location choices because job listings are classified by area code. We used county-level data to generate area code–level data for analysis (procedure discussed below). Three kinds of county-level data are used: county-level poverty/income data, county-level current business patterns (CBP) data, and county-level crime data. The first two datasets were downloaded from the US Census Bureau. County-level crime data was obtained from the US Department of Justice.5 Compared to other data sources such as American Community Survey (ACS), our dataset has several advantages. First, ACS data can provide the distribution of Chinese immigrants (or Chinese Americans) in different types of destinations, but the ACS cannot identify whether Chinese immigrants work for Chinese restaurant/business owners. We think this information is important to link to the immigrant enclave thesis with an understanding of the immigrant labor market. Second, ACS data would not allow the story of settlement mechanisms to new destinations to be explored. The use of our own survey and existing data along with a discussion of EAs and Chinatown buses is critical to unveiling the story of this new pattern of Chinese immigrant settlement. Third, we note that ACS data is limited to self-reported salary while our survey contains salary information provided by employers. The salary level in our data is set by employers and is not a direct function of individual worker–level characteristics. Salary set by employers instead of self-reported salary should, therefore, minimize selection bias. It is possible that Chinese immigrants with characteristics positively associated with earnings, such as education level, ambition, or motivation, are more likely to move to new destinations, and using the ACS to observe salary differences between Chinese immigrants who reside in new immigrant destinations versus traditional destinations does not effectively account for these characteristics. Analytic Strategy (1) Mapping of Business Locations and Construction of Area Code–Level Data Our survey resulted in a list of jobs identified by telephone area codes. Using job counts of Chinese restaurants located in a specific area code, we mapped the distribution of business locations by area code using an area code boundary file. Both area code boundary and county-level boundary files are available from ESRI Data and Maps.6 We used both these files to map patterns of job distribution by other characteristics at the area code level (i.e., median household income, poverty level, and crime rate). To do this, we “converted” county-level information to the area code level (using ArcMap’s clip, merge, and dissolve features) by overlaying the area code boundaries with the county boundaries. In this manner, we aggregated the county-level data into area code–level attributes. For example, if, hypothetically, the area code 555 roughly corresponds to the combination of two adjoining counties A and B, then we sum the corresponding (i.e., A and B) counties’ attributes, such as population size or number of businesses. The summed totals become the attributes for area code 555. However, it is also often the case that a county polygon is not entirely contained within a single area code. For example, county C may contain area codes 555 and 666. In this case, we proportionally split the attributes (such as population size) of county C based on the spatial distribution of each area code 555 and 666 within county C. Whether combining or splitting the county attributes, this process ultimately created our area code–level data. To assess the potential impact of imprecision between area codes and the counties whose attributes were mapped onto the area codes, we created an indicator (dummy) variable, scored 1 if the area code polygon fully enveloped its component counties, and 0 otherwise, and included this as a control variable in each of our regression models. None of the dummy variables was statistically significant. We then refit the regressions without the dummy variable and compared the estimated coefficients, standard errors, and test statistics. None of the statistical inference differed between models with and without the dummy variable, so we report the models without the dummy control for area code to component-county mismatch. (2) Statistical Models of Location Choices of Chinese Restaurants Using Area Code Data Following the procedure described above, we obtain an area code–level data file with the distribution of Chinese restaurant jobs as well as basic economic, business, and crime data at the area code level. Our outcome of interest is the number of jobs in an area code zone. However, to use information from both (2010 and 2011) surveys, we averaged the number of jobs in each area code over two years. The resulting averaged counts are not always integers; however, the number of jobs can reasonably be assumed to be randomly drawn from each area code’s Poisson distribution with a single parameter, λ, determining the properties of that distribution. This parameter, λ, may be interpreted as an area code’s job listing rate of occurrence or propensity over a given period (Berk and MacDonald 2007; Long 1997; Long and Freese 2006). In addition, our data collection of job listings likely involves noisy measurement that introduces random variation into λi ⁠, thereby altering the Poisson formulation by building in a second source of uncertainty (Berk and MacDonald 2007, 13; Greene 2003, 744–45). This leads us to utilize the negative binomial distribution. While we agree with Berk and MacDonald that the negative binomial regression model (NBRM) is overused for count data and misunderstood as an all-purpose “fix” to excess variation in the conditional distribution of count data, our application appears to be justified by an expectation of excess variation in the stochastic part of the model. Accordingly, we can write the NBRM with the rate, μi ⁠, as a function of observed xks and a disturbance term (as in normal regression) that reflects unobserved heterogeneity among the observations: lnμi=xiTβ+εi Exponentiating both sides yields μi=exp(xiTβ)exp(εi)=exp(xiTβ)δi, where we define δ≡exp(ε) and then assume that δ is drawn from a gamma distribution and normalized so that the expected value E(δ)=1 (Long and Freese 2006, 372–73; Berk and MacDonald 2007, 12–14). Our NBRM draws on several literatures to specify the systematic part of the model. Drawing from the economic sociology of immigration (Portes 1995), we expect that higher counts of job listings are more likely to be in Asian/Chinese immigrant–concentrated places; therefore, our model includes measures for the proportion of total population that is non-Hispanic white, non-Hispanic black, and Asian. As mentioned earlier, we are not able to specifically measure the Chinese proportion of total population in each area code, so we use the Asian proportion as a close proxy. The “transition zones” perspective from the human ecological theory predicts that immigrant businesses and jobs are located in transition zones with high social disorganization and crime rates. Alternatively, the “business climate” perspective advocated by economists suggests that immigrant businesses and jobs are more likely located in places with low unemployment, a low crime rate, and an established business climate (Card and Lewis 2005). Accordingly, our regressions include median household income, unemployment rate, property crime rate, and the number of businesses in the accommodation and food sectors. We also control for population size (population and population-squared) and geographic size of the area code zone (square miles). Our decision to allow for nonlinearity between the number of jobs and population size is motivated by our pre-regression observation of a positive association that holds until the area code population size surpasses four million, then the association becomes negative. Although there are not many such large population cases and our ability to theorize about them is not sufficiently developed, there are enough such cases to suggest that the negative association is real and that our model specification should allow for this functional form. We also consider that controlling for the size of area codes is not sufficient to eliminate the impact of spatial heterogeneity. Indeed, map 2 reveals the large disparity in the size of US area codes. Many states in the North Central, West, and Southwest regions are composed of one or several area codes, while states in the East, South, and Central regions are highly subdivided. Studies have documented spatial dependence between areal units in regression-based analyses. For example, Crowder and South (2008) considered the influence of extra-local conditions on white flight in the context of residential mobility research. Both extra-local conditions and economic activity in any region that borders the focal one may impact the conditions within the focal unit. Of primary concern to this study, the number of job listings in an area code may be correlated with job listings in nearby or adjacent area codes because immigrant entrepreneurs may open restaurants near other successful restaurants. Even under the best circumstances, a maximum-likelihood (ML) estimator of spatial dependence is not well suited to probability models, including models for count data, and when coupled with severe disparity in the size of areal units, the conditions necessary for model convergence are rarely met. Map 2 View largeDownload slide Spatial distribution of jobs by phone area codes. Note: The shade indicates number of jobs in each area code, with darker shade showing a larger number of jobs in the area code. Map 2 View largeDownload slide Spatial distribution of jobs by phone area codes. Note: The shade indicates number of jobs in each area code, with darker shade showing a larger number of jobs in the area code. We address the potential for spatial heterogeneity and a specific form for spatial dependence by including state fixed effects (a vector of dummy variables that uniquely identify each state). Our approach is motivated by the observation that highly subdivided states will have more within-state than between-state area code “neighbors,” while states with only one or few area codes have such a large area code size that between-state spatial dependence is an unlikely consequence. (3) Supplementary Models of the Earnings of Immigrant Workers To examine how spatial location of jobs relates to wages for immigrant workers, we estimate two additional types of models. The first uses jobs as the unit of analysis and estimates multilevel models using information about jobs as the first level and information about job location (area code) as the second level (Raudenbush and Bryk 2002). The advantage of this approach is that we estimate how certain job characteristics are related to wages, in addition to how our key interest in distance from NYC is related to wages. Since we are using characteristics for each job, we can use either our 2010 or 2011 survey of EAs. We decided to use the 2011 survey for this purpose.7 Our second strategy estimates conventional spatial lag and spatial error regression models using area code as the unit of analysis and the mean monthly wage at the area code level as our dependent variable (again using the 2011 EA survey data) and other area code–level variables as predictor variables (Anselin 1998; Chi and Zhu 2008). Following Anselin (1998) and Chi and Zhu (2008), we estimate the following spatial lag model with a “queen” contiguity spatial weights matrix: Y=Xβ+λWY+ε where Y denotes the logged mean monthly salary for jobs listed in each area code, X represents a matrix of exogenous explanatory variables at the area code level, WY introduces endogeneity in the outcome by connecting neighboring area codes, and ε is the error term. We also estimate a spatial error model in the following form: Y=Xβ+εandε=λWε+μ where Wε allows for unobserved correlated effects that are quantified in λ ⁠. For all wage-related spatial models, we control for cost of living along with selected area code–level characteristics. Results Table 1 provides descriptive information about variables used in our analysis. Table 2a presents the distribution of number of jobs in each area code from our wave 1 survey in 2010. Overall, we see that Chinese restaurant jobs spread into 64 percent of the 274 area codes in the United States. Nearly 24 percent and 20 percent of area codes have 1–4 and 5–14 jobs listed for Chinese restaurants, respectively. Over 20 percent of area code zones are characterized by high-density job availability, defined as a minimum of 15 jobs in each zone. Similar results are reported in table 2b using the 2011 wave 2 data. Table 1. Descriptive Statistics of Variables Used in the Analysis Variable Mean Std. dev Number of jobs in area code (2010) 7.84 11.65 Number of jobs in area code (2011) 8.42 11.42 Percentage of non-Hispanic white (2010) 65.56 18.60 Percentage of black (2010) 12.71 10.66 Percentage of Asian (2010) 4.35 4.73 Population (2010) 1,469,162.00 1,065,010.00 % 2010 median household income <$40,000 9.52 29.41 % 2010 median household income range $40,000–$50,000 40.66 49.21 % 2010 median household income range $50,000–$60,000 27.84 44.90 % 2010 median household income range $60,000–$70,000 14.65 35.43 % 2010 median household income range ≥$70,000 7.33 26.10 Percentage of unemployment population (2005) 4.75 0.93 Property crime rate (2010) 0.02 0.01 Number of establishments (accommodation & food service) 3016.08 2115.15 Cost of living (2010) 930.37 282.34 Variable Mean Std. dev Number of jobs in area code (2010) 7.84 11.65 Number of jobs in area code (2011) 8.42 11.42 Percentage of non-Hispanic white (2010) 65.56 18.60 Percentage of black (2010) 12.71 10.66 Percentage of Asian (2010) 4.35 4.73 Population (2010) 1,469,162.00 1,065,010.00 % 2010 median household income <$40,000 9.52 29.41 % 2010 median household income range $40,000–$50,000 40.66 49.21 % 2010 median household income range $50,000–$60,000 27.84 44.90 % 2010 median household income range $60,000–$70,000 14.65 35.43 % 2010 median household income range ≥$70,000 7.33 26.10 Percentage of unemployment population (2005) 4.75 0.93 Property crime rate (2010) 0.02 0.01 Number of establishments (accommodation & food service) 3016.08 2115.15 Cost of living (2010) 930.37 282.34 Sources: New York City Chinatown Employment Agency Survey 2010; ICPSR 20660: County Characteristics, 2000–2007 [United States]; 2009 County Business Patterns (CBP). Table 1. Descriptive Statistics of Variables Used in the Analysis Variable Mean Std. dev Number of jobs in area code (2010) 7.84 11.65 Number of jobs in area code (2011) 8.42 11.42 Percentage of non-Hispanic white (2010) 65.56 18.60 Percentage of black (2010) 12.71 10.66 Percentage of Asian (2010) 4.35 4.73 Population (2010) 1,469,162.00 1,065,010.00 % 2010 median household income <$40,000 9.52 29.41 % 2010 median household income range $40,000–$50,000 40.66 49.21 % 2010 median household income range $50,000–$60,000 27.84 44.90 % 2010 median household income range $60,000–$70,000 14.65 35.43 % 2010 median household income range ≥$70,000 7.33 26.10 Percentage of unemployment population (2005) 4.75 0.93 Property crime rate (2010) 0.02 0.01 Number of establishments (accommodation & food service) 3016.08 2115.15 Cost of living (2010) 930.37 282.34 Variable Mean Std. dev Number of jobs in area code (2010) 7.84 11.65 Number of jobs in area code (2011) 8.42 11.42 Percentage of non-Hispanic white (2010) 65.56 18.60 Percentage of black (2010) 12.71 10.66 Percentage of Asian (2010) 4.35 4.73 Population (2010) 1,469,162.00 1,065,010.00 % 2010 median household income <$40,000 9.52 29.41 % 2010 median household income range $40,000–$50,000 40.66 49.21 % 2010 median household income range $50,000–$60,000 27.84 44.90 % 2010 median household income range $60,000–$70,000 14.65 35.43 % 2010 median household income range ≥$70,000 7.33 26.10 Percentage of unemployment population (2005) 4.75 0.93 Property crime rate (2010) 0.02 0.01 Number of establishments (accommodation & food service) 3016.08 2115.15 Cost of living (2010) 930.37 282.34 Sources: New York City Chinatown Employment Agency Survey 2010; ICPSR 20660: County Characteristics, 2000–2007 [United States]; 2009 County Business Patterns (CBP). Table 2a. Distribution of Jobs at the Area Code Level (Survey 1: 2010) Number of jobs in the area code Frequency Percent 0 96 35.04 1–4 66 24.09 5–14 55 20.07 15–24 31 11.31 25–34 16 5.84 35+ 10 3.65 Total 274 100.00 Number of jobs in the area code Frequency Percent 0 96 35.04 1–4 66 24.09 5–14 55 20.07 15–24 31 11.31 25–34 16 5.84 35+ 10 3.65 Total 274 100.00 Source: New York City Chinatown Employment Agency Survey 2010 (Total number of jobs = 2147). Table 2a. Distribution of Jobs at the Area Code Level (Survey 1: 2010) Number of jobs in the area code Frequency Percent 0 96 35.04 1–4 66 24.09 5–14 55 20.07 15–24 31 11.31 25–34 16 5.84 35+ 10 3.65 Total 274 100.00 Number of jobs in the area code Frequency Percent 0 96 35.04 1–4 66 24.09 5–14 55 20.07 15–24 31 11.31 25–34 16 5.84 35+ 10 3.65 Total 274 100.00 Source: New York City Chinatown Employment Agency Survey 2010 (Total number of jobs = 2147). Table 2b. Distribution of Jobs at the Area Code Level (Survey 2: 2011) Number of jobs in the area code Frequency Percent 0 74 27.01 1–4 75 27.37 5–14 64 23.36 15–24 36 13.14 25–34 17 6.20 35+ 8 2.92 Total 274 100.00 Number of jobs in the area code Frequency Percent 0 74 27.01 1–4 75 27.37 5–14 64 23.36 15–24 36 13.14 25–34 17 6.20 35+ 8 2.92 Total 274 100.00 Source: New York City Chinatown Employment Agency Survey 2011 (Total number of jobs = 2316). Table 2b. Distribution of Jobs at the Area Code Level (Survey 2: 2011) Number of jobs in the area code Frequency Percent 0 74 27.01 1–4 75 27.37 5–14 64 23.36 15–24 36 13.14 25–34 17 6.20 35+ 8 2.92 Total 274 100.00 Number of jobs in the area code Frequency Percent 0 74 27.01 1–4 75 27.37 5–14 64 23.36 15–24 36 13.14 25–34 17 6.20 35+ 8 2.92 Total 274 100.00 Source: New York City Chinatown Employment Agency Survey 2011 (Total number of jobs = 2316). Map 2 shows the distribution of jobs obtained from our EA survey in NYC. As we can see, the job distribution clearly spreads widely across the United States. We identify job frequency using different colors, with red indicating the largest number of available jobs, followed by blue, green, and orange. Overall, these job listings spread across 37 states. There is a striking spatial pattern of job distribution. The density of available jobs declines with distance from NYC, and there are few jobs located near the West Coast, which seems to be a separate labor market for immigrants. As such, although the labor market for NYC-based immigrants is largely national, it does not quite extend to the West Coast. Map 2 is also consistent with the classification method proposed by Massey (2008): big five, second tier, new destinations, and remaining states. With Massey’s classification scheme, jobs in new destinations and remaining states represent 55 percent of the job listings. This reveals a dramatic geographic diversification of restaurant job locations and high concentration of jobs in non-gateway destinations in the United States. One advantage of working in traditional destinations for immigrants is that these locations tend to be in large cities with easy access to transportation. However, with large numbers of newly arrived immigrants working in faraway locations and rural towns, transportation can be a challenge. Since 1997, Chinese immigrant entrepreneurs have been quick to start new long-distance bus companies to serve the needs of immigrants (NYC Department of City Planning 2009). While conducting fieldwork on EAs, we found large stacks of business cards from bus companies in Chinatown. We systematically collected all bus information from each EA. There are two major bus services. One is mainly for short distances (to Boston or Philadelphia), and the other is for long distances. We identified three major long-distance bus routes that serve the needs of immigrant workers. The first route goes along the East Coast, with Orlando as the final stop. The second bus route goes to the Midwest, with Kansas City as the destination. The third bus line goes to the Deep South, with three possible destinations of Alabama, Mississippi, and Tennessee. We conducted an analysis of the determinants of job distributions by area code. We used the following area code–level characteristics: median household income, unemployment rate, crime rate, proportion of non-Hispanic whites and the proportion of blacks, and number of business establishments providing accommodations and food. Because our dependent variable is the number of jobs in each area code, we estimated a negative binomial model. For this model, state-level fixed effects are included in model B. Results are shown in table 3. We note that Moran’s I, a global measure of spatial autocorrelation, for model A is statistically significant, but once we enter state-level fixed effects in model B, Moran’s I becomes insignificant, indicating that spatial dependence has been purged from the model. Table 3. Estimated Coefficients from Negative Binomial Models of Number of Jobs in Area Code Zone Model A Model B* Independent variables B SE B SE Intercept −6.7407*** 0.7263 −27.1529*** 1.7491 Non-Hispanic white percentage 0.0402*** 0.0074 0.0274** 0.0093 Black percentage 0.0486*** 0.0109 −0.0018 0.0149 Asian percentage −0.2178*** 0.0310 −0.2393*** 0.0554 Population in the area code (in 100,000) 0.0539 0.0283 0.1397** 0.0438 Population in the area code (in 100,000) squared −0.0010* 0.0004 −0.0013* 0.0005 Square miles of the area code (in 10,000) 0.1726* 0.0757 0.3685** 0.1131 Median household income  $40,000 or below 1.5366* 0.6299  $40,000–50,000 1.4818** 0.5434  $50,000–60,000 1.2500* 0.5358  $60,000–70,000 0.3491 0.4688  $70,000 or above (reference) – – Unemployment proportion −0.7654*** 0.2290 Property crime rate −0.0001* 0.0000 Number of establishments in 1,000 (food &accommodation) −0.1537 0.1481 z-ratio of Moran’s I statistic 11.166*** 1.563 Log likelihood 4685.0115 4766.2909 AIC 1664.9960 1612.4373 Model A Model B* Independent variables B SE B SE Intercept −6.7407*** 0.7263 −27.1529*** 1.7491 Non-Hispanic white percentage 0.0402*** 0.0074 0.0274** 0.0093 Black percentage 0.0486*** 0.0109 −0.0018 0.0149 Asian percentage −0.2178*** 0.0310 −0.2393*** 0.0554 Population in the area code (in 100,000) 0.0539 0.0283 0.1397** 0.0438 Population in the area code (in 100,000) squared −0.0010* 0.0004 −0.0013* 0.0005 Square miles of the area code (in 10,000) 0.1726* 0.0757 0.3685** 0.1131 Median household income  $40,000 or below 1.5366* 0.6299  $40,000–50,000 1.4818** 0.5434  $50,000–60,000 1.2500* 0.5358  $60,000–70,000 0.3491 0.4688  $70,000 or above (reference) – – Unemployment proportion −0.7654*** 0.2290 Property crime rate −0.0001* 0.0000 Number of establishments in 1,000 (food &accommodation) −0.1537 0.1481 z-ratio of Moran’s I statistic 11.166*** 1.563 Log likelihood 4685.0115 4766.2909 AIC 1664.9960 1612.4373 Note: *p < 0.05 **p < 0.01 ***p < 0.001 Table 3. Estimated Coefficients from Negative Binomial Models of Number of Jobs in Area Code Zone Model A Model B* Independent variables B SE B SE Intercept −6.7407*** 0.7263 −27.1529*** 1.7491 Non-Hispanic white percentage 0.0402*** 0.0074 0.0274** 0.0093 Black percentage 0.0486*** 0.0109 −0.0018 0.0149 Asian percentage −0.2178*** 0.0310 −0.2393*** 0.0554 Population in the area code (in 100,000) 0.0539 0.0283 0.1397** 0.0438 Population in the area code (in 100,000) squared −0.0010* 0.0004 −0.0013* 0.0005 Square miles of the area code (in 10,000) 0.1726* 0.0757 0.3685** 0.1131 Median household income  $40,000 or below 1.5366* 0.6299  $40,000–50,000 1.4818** 0.5434  $50,000–60,000 1.2500* 0.5358  $60,000–70,000 0.3491 0.4688  $70,000 or above (reference) – – Unemployment proportion −0.7654*** 0.2290 Property crime rate −0.0001* 0.0000 Number of establishments in 1,000 (food &accommodation) −0.1537 0.1481 z-ratio of Moran’s I statistic 11.166*** 1.563 Log likelihood 4685.0115 4766.2909 AIC 1664.9960 1612.4373 Model A Model B* Independent variables B SE B SE Intercept −6.7407*** 0.7263 −27.1529*** 1.7491 Non-Hispanic white percentage 0.0402*** 0.0074 0.0274** 0.0093 Black percentage 0.0486*** 0.0109 −0.0018 0.0149 Asian percentage −0.2178*** 0.0310 −0.2393*** 0.0554 Population in the area code (in 100,000) 0.0539 0.0283 0.1397** 0.0438 Population in the area code (in 100,000) squared −0.0010* 0.0004 −0.0013* 0.0005 Square miles of the area code (in 10,000) 0.1726* 0.0757 0.3685** 0.1131 Median household income  $40,000 or below 1.5366* 0.6299  $40,000–50,000 1.4818** 0.5434  $50,000–60,000 1.2500* 0.5358  $60,000–70,000 0.3491 0.4688  $70,000 or above (reference) – – Unemployment proportion −0.7654*** 0.2290 Property crime rate −0.0001* 0.0000 Number of establishments in 1,000 (food &accommodation) −0.1537 0.1481 z-ratio of Moran’s I statistic 11.166*** 1.563 Log likelihood 4685.0115 4766.2909 AIC 1664.9960 1612.4373 Note: *p < 0.05 **p < 0.01 ***p < 0.001 Predictions from the economic sociology of immigration and migration networks perspective do not fare well. Chinese restaurants and jobs generally are in places with high rates of non-Hispanic whites and blacks, but not in places with a high concentration of Asians/Chinese. This is consistent with our interviews in which immigrant entrepreneurs call all locations not in NYC “waizhou” (outside New York State). “Waizhou” connotes being far away from the Chinese immigrant community. In line with the business climate perspective, these restaurant jobs tend to be in areas with low unemployment rates. Restaurant owners tend to start businesses in low-crime areas; the coefficient for crime rate is negative and statistically significant. Again, from our interviews, restaurant entrepreneurs often told us that when opening businesses, they look for “hao qu” (good community or neighborhood), which could be interpreted as low unemployment and a low crime rate. Some educated business owners use the Internet to find out this information. For example, Mr. Zheng, an immigrant restaurant owner in Ohio, used the Internet to research the income level of the neighborhood when he was planning his restaurant’s location. Other immigrant entrepreneurs obtained local information from real estate brokers. The low crime rate is especially important for these owners because there have been frequent reports of tragedies for those restaurants that are often located in poor neighborhoods. We also note that coefficients for low- or middle-income variables are statistically significant. Meals in Chinese restaurants are very affordable, especially at Chinese takeout places. As such, potential customers do not need to earn a high income to eat in a Chinese restaurant. But the key to a consistent customer base is steady employment within the restaurants’ neighborhood, which provides a stable income. Therefore, Chinese restaurant owners are sensitive to unemployment in a community. We next assess how salary is related to new destinations. The substantive question is “Do immigrants receive higher compensation once they move out of NYC?” The focal covariate that operationalizes this research question is a distance measure based on a centroid point in each location between Manhattan (212 area code) to each destination area code. Table 4 reports results from the multilevel models of log monthly salary for migrant workers. The “distance from NYC” coefficient, in model A and model B, is positive and statistically significant: as jobs move farther away from Manhattan, the expected monthly salary increases. Model B further disaggregates the log monthly salary regression by job category (cook [the reference category], waiter/waitress, takeout delivery, and others). Surprisingly, waiters/waitresses are paid higher salaries than cooks. We speculate that this is because waiters/waitress require English-language skills to communicate with dine-in customers and to take orders over the phone. Among recently arrived, low-skilled Chinese immigrants (often with a middle school education), this level of English proficiency is not universal and thus is rewarded. In contrast, getting training and experience in cooking skills is relatively easy. Table 4. Estimated Coefficients from Multilevel Model of Monthly Salary(logged) Model A Model B Independent variables B SE B SE Intercept 7.8633*** 0.0865 7.9537*** 0.0732 Job-level characteristics  Job category   Cook (reference)   Waiter/waitress – 0.0307*** 0.0077   Takeout – −0.0131 0.0132   Others – −0.1848*** 0.0069 Area code–level characteristics  Distance from NYC 0.0112*** 0.0012 0.0084*** 0.0010  Asian percentage −0.0034 0.0024 −0.0026 0.0020  Population in the area code (in 100,000) 0.0010 0.0006 0.0010* 0.0005  Square miles of the area code (in 10,000) −0.0096 0.0068 −0.0056 0.0058  Percent of population in labor force −0.0962 0.1504 −0.1982 0.1267  Cost of living/100 −0.0044 0.0028 −0.0021 0.0023  Unemployment proportion 0.0088 0.0066 0.0074 0.0056 Model A Model B Independent variables B SE B SE Intercept 7.8633*** 0.0865 7.9537*** 0.0732 Job-level characteristics  Job category   Cook (reference)   Waiter/waitress – 0.0307*** 0.0077   Takeout – −0.0131 0.0132   Others – −0.1848*** 0.0069 Area code–level characteristics  Distance from NYC 0.0112*** 0.0012 0.0084*** 0.0010  Asian percentage −0.0034 0.0024 −0.0026 0.0020  Population in the area code (in 100,000) 0.0010 0.0006 0.0010* 0.0005  Square miles of the area code (in 10,000) −0.0096 0.0068 −0.0056 0.0058  Percent of population in labor force −0.0962 0.1504 −0.1982 0.1267  Cost of living/100 −0.0044 0.0028 −0.0021 0.0023  Unemployment proportion 0.0088 0.0066 0.0074 0.0056 Note: *p < 0.05 **p < 0.01 ***p < 0.001 Sources: The 2011 New York City Chinatown Employment Agency Survey; http://censtats.census.gov/usa/usa.shtml: 2010 U.S. County Characteristics. Table 4. Estimated Coefficients from Multilevel Model of Monthly Salary(logged) Model A Model B Independent variables B SE B SE Intercept 7.8633*** 0.0865 7.9537*** 0.0732 Job-level characteristics  Job category   Cook (reference)   Waiter/waitress – 0.0307*** 0.0077   Takeout – −0.0131 0.0132   Others – −0.1848*** 0.0069 Area code–level characteristics  Distance from NYC 0.0112*** 0.0012 0.0084*** 0.0010  Asian percentage −0.0034 0.0024 −0.0026 0.0020  Population in the area code (in 100,000) 0.0010 0.0006 0.0010* 0.0005  Square miles of the area code (in 10,000) −0.0096 0.0068 −0.0056 0.0058  Percent of population in labor force −0.0962 0.1504 −0.1982 0.1267  Cost of living/100 −0.0044 0.0028 −0.0021 0.0023  Unemployment proportion 0.0088 0.0066 0.0074 0.0056 Model A Model B Independent variables B SE B SE Intercept 7.8633*** 0.0865 7.9537*** 0.0732 Job-level characteristics  Job category   Cook (reference)   Waiter/waitress – 0.0307*** 0.0077   Takeout – −0.0131 0.0132   Others – −0.1848*** 0.0069 Area code–level characteristics  Distance from NYC 0.0112*** 0.0012 0.0084*** 0.0010  Asian percentage −0.0034 0.0024 −0.0026 0.0020  Population in the area code (in 100,000) 0.0010 0.0006 0.0010* 0.0005  Square miles of the area code (in 10,000) −0.0096 0.0068 −0.0056 0.0058  Percent of population in labor force −0.0962 0.1504 −0.1982 0.1267  Cost of living/100 −0.0044 0.0028 −0.0021 0.0023  Unemployment proportion 0.0088 0.0066 0.0074 0.0056 Note: *p < 0.05 **p < 0.01 ***p < 0.001 Sources: The 2011 New York City Chinatown Employment Agency Survey; http://censtats.census.gov/usa/usa.shtml: 2010 U.S. County Characteristics. In table 5, we estimate spatial models of monthly salary at the area code level. Again, results from the spatial error (model A) and spatial lag (model B) models support the inference that distance from NYC increases log monthly salary. We note that although this finding of a more favorable salary for Chinese who live away from traditional immigrant destinations is consistent with earlier studies using ACS data (see Flippen and Kim 2015), the potential individual selectivity bias is minimized in our case. It is possible that Chinese immigrants with characteristics positively associated with earnings, such as education level, ambition, or motivation, are more likely to move to new destinations, and using the ASC to observe salary differences between Chinese immigrants who reside in new immigrant destinations versus traditional destinations does not account for these characteristics. The salary level in our data is set by employers and is not a direct function of individual worker–level characteristics. Table 5. Estimated Coefficients from Spatial Models of Salary (logged) at the Area Code Level Model A (spatial error model) Model B (spatial lag model) Independent variables B SE B SE Intercept 7.9525*** 0.1151 7.9497*** 0.1155 Distance from NYC 0.0001*** 0.00002 0.0001*** 0.00002 Asian percentage −0.0053 0.0030 −0.0054 0.0030 Population in the area code (in 100,000) 0.0004 0.0006 0.0004 0.0006 Square miles of the area code (in 10,000) −0.0085 0.0047 −0.0084 0.0048 Percent of population in labor force −0.1556 0.1683 −0.1549 0.1683 Cost of living/100 −0.0008 0.0041 −0.0007 0.0041 Unemployment percentage −0.0021 0.0070 −0.0021 0.0070 Lambda 0.00002 0.00007 Rho 0.00002 0.00007 Log likelihood 226.6569 226.6686 z-ratio of Moran’s I statistic 1.399 1.399 Model A (spatial error model) Model B (spatial lag model) Independent variables B SE B SE Intercept 7.9525*** 0.1151 7.9497*** 0.1155 Distance from NYC 0.0001*** 0.00002 0.0001*** 0.00002 Asian percentage −0.0053 0.0030 −0.0054 0.0030 Population in the area code (in 100,000) 0.0004 0.0006 0.0004 0.0006 Square miles of the area code (in 10,000) −0.0085 0.0047 −0.0084 0.0048 Percent of population in labor force −0.1556 0.1683 −0.1549 0.1683 Cost of living/100 −0.0008 0.0041 −0.0007 0.0041 Unemployment percentage −0.0021 0.0070 −0.0021 0.0070 Lambda 0.00002 0.00007 Rho 0.00002 0.00007 Log likelihood 226.6569 226.6686 z-ratio of Moran’s I statistic 1.399 1.399 Note: *p < 0.05 **p < 0.01 ***p < 0.001 Sources: The 2011 New York City Chinatown Employment Agency Survey; http://censtats.census.gov/usa/usa.shtml: 2010 U.S. County Characteristics. Table 5. Estimated Coefficients from Spatial Models of Salary (logged) at the Area Code Level Model A (spatial error model) Model B (spatial lag model) Independent variables B SE B SE Intercept 7.9525*** 0.1151 7.9497*** 0.1155 Distance from NYC 0.0001*** 0.00002 0.0001*** 0.00002 Asian percentage −0.0053 0.0030 −0.0054 0.0030 Population in the area code (in 100,000) 0.0004 0.0006 0.0004 0.0006 Square miles of the area code (in 10,000) −0.0085 0.0047 −0.0084 0.0048 Percent of population in labor force −0.1556 0.1683 −0.1549 0.1683 Cost of living/100 −0.0008 0.0041 −0.0007 0.0041 Unemployment percentage −0.0021 0.0070 −0.0021 0.0070 Lambda 0.00002 0.00007 Rho 0.00002 0.00007 Log likelihood 226.6569 226.6686 z-ratio of Moran’s I statistic 1.399 1.399 Model A (spatial error model) Model B (spatial lag model) Independent variables B SE B SE Intercept 7.9525*** 0.1151 7.9497*** 0.1155 Distance from NYC 0.0001*** 0.00002 0.0001*** 0.00002 Asian percentage −0.0053 0.0030 −0.0054 0.0030 Population in the area code (in 100,000) 0.0004 0.0006 0.0004 0.0006 Square miles of the area code (in 10,000) −0.0085 0.0047 −0.0084 0.0048 Percent of population in labor force −0.1556 0.1683 −0.1549 0.1683 Cost of living/100 −0.0008 0.0041 −0.0007 0.0041 Unemployment percentage −0.0021 0.0070 −0.0021 0.0070 Lambda 0.00002 0.00007 Rho 0.00002 0.00007 Log likelihood 226.6569 226.6686 z-ratio of Moran’s I statistic 1.399 1.399 Note: *p < 0.05 **p < 0.01 ***p < 0.001 Sources: The 2011 New York City Chinatown Employment Agency Survey; http://censtats.census.gov/usa/usa.shtml: 2010 U.S. County Characteristics. Summary and Conclusion This paper aims to examine a new employment and settlement pattern for low-skilled Chinese immigrants in the United States. Relying on our unique surveys of EAs in NYC’s Chinatown, we systematically analyze this new pattern. Our results show that there has been a fundamental shift in patterns of settlement among recently arrived, low-skilled Chinese immigrants. The days of these immigrants congregating exclusively in Chinatown for jobs and settlement are long gone. A substantial number of low-skilled jobs today are located across the United States. This decisive shift in employment patterns of low-skilled Chinese immigrants is due to several critical factors. First, the broad economic picture is that NYC’s market has made it difficult and unappealing to open new Chinese restaurants. After all, there is a limited customer base and the high commercial rent prompts immigrant restaurant entrepreneurs to look beyond the city. Although our fieldwork suggests that there have always been Chinese immigrants who ventured outside New York City and opened restaurants in remote places, nothing compares to the scale of present-day expansion.8 The Chinese restaurant business today has reached almost everywhere: from Alabama to Alaska. Thus, the title of our paper, “From Chinatown to Every Town,” is not an exaggeration. Beyond factors related to NYC’s restaurant market, we argue that two more factors were critically important in facilitating this expansion: the role of EAs and Chinatown buses. EAs established market mechanisms to build a bridge between employers and potential employees. The Chinatown buses greatly reduced the transportation needs of low-skilled Chinese immigrants. With convenient routes and reasonable prices, Chinese immigrants can easily navigate across geographic spaces without a high degree of English proficiency. The fundamental shift to non-gateway destinations has several important consequences. The trend of moving to non-gateway destinations opens up a broader labor market that can provide employment opportunities for a substantially larger number of immigrants. In this sense, the recent significant entry of immigrants from Fujian province cannot be sustained without these Chinese immigrant entrepreneurs opening restaurants in different parts of the country. Likewise, without the large supply of immigrant workers, these entrepreneurs cannot realize their American dream of starting their own business. This is a win-win situation for entrepreneurs and low-skilled immigrant workers. It is often the case that geographic mobility leads to socioeconomic mobility. Immigrants who choose to work in non-gateway destinations can reap a significantly higher salary than immigrant workers in NYC. The current study echoes studies that find that immigrants working outside ethnic enclaves obtain better economic advancement than those working in the enclaves (Sanders and Nee 1987; Xie and Gough 2011). Recent literature seems to give short shrift to the issue of how new destinations are linked to socioeconomic mobility for immigrants. Our study provides evidence that moving out in this case means moving up, as measured by increased salary. Previous scholars reveal that this new settlement pattern in non-gateway destinations leads to new challenges such as the increase of conflicts for race and ethnic relations, especially in locations where residents have not encountered large numbers of immigrants before (Fennelly 2005; Marrow 2011; Massey and Capoferro 2008). However, our fieldwork in Chinatown reveals that this new settlement pattern may be improving some inter-group relations. Because of the short supply of Chinese immigrant workers in non-gateway locations, entrepreneurs often rely on non-Chinese workers, especially Latino workers. To accommodate this link, there is an EA that specializes in recruiting Latino workers for Chinese entrepreneurs. To attract Latino workers, staff members in the agency display currencies from major Latin American countries, which creates a familiar and comfortable environment for the job search process. Staff members also facilitate the job search process by knowing simple Spanish phases to ensure communication with potential workers. While this certainly provides employment opportunities for Latino workers, this process raises interesting research questions as well. For example, if we follow the “training system” logic of Bailey and Waldinger (1991), Chinese immigrants working for Chinese employers are likely to become entrepreneurs themselves, thus leading to social mobility. But it is unknown whether this training system extends to the Latino workers in Chinese immigrant–owned businesses. Turning to the determinants of location choices of Chinese restaurants and jobs, we find that the economic sociology of immigration that links ethnic businesses near immigrant residential locations was not supported. In fact, Chinese restaurant locations and jobs are strongly associated with the presence of non-Hispanic whites. This implies that Chinese business owners cater to a non-Hispanic white customer base, not other Chinese customers. This challenges the traditional wisdom that ethnic-owned businesses need to cater to other ethnic group members (Bonacich and Modell 1980; Light 2006; Portes 1995; Portes and Rumbaut 2006). Our study also sheds light on why the expansion of Chinese restaurant businesses into vast areas in the country is so successful. Chinese restaurant owners seem to focus on attracting relatively low- or middle-income groups as their main patrons (see the significant coefficients for low- or middle-income variables for table 3). This strategy ultimately ensures a large potential customer market. Theoretically, our study has implications for a need to (re)conceptualize the immigrant labor market (Sassen 1995). Much of the existing literature tends to conceptualize the immigrant labor market as being within a city or metropolitan area and studies how immigrants choose to work in an ethnic enclave or outside it or in a suburban area (see Fong, Luk, and Ooka 2005). As Fong, Luk, and Ooka (2005) argued, the human ecological perspective, based on the early twentieth-century experiences of European immigrants typically concentrated in central cities, does not explain the current spatial patterns of Chinese businesses in suburban Toronto. In the case of low-skilled Chinese immigrants in the United States, this somewhat narrow conceptualization of the immigrant labor market is clearly not sufficient, as many Chinese immigrant workers choose to work in faraway places across the country. To study spatial locations of current immigrant businesses and workers, our paper calls for broadening the understanding of the immigrant labor market to a wider geographic region in this country. This is true not only for Chinese immigrants but also for Mexican immigrants, as Massey, Durand, and Pren (2016) revealed recently that the spatial settlement pattern has now shifted from three states to 50 states in the United States. The implications of our study go beyond sociology of immigration. A well-known economist, Enrico Moretti (2012), wrote recently: “[the] job market for professional positions is a national one, while the job market for manual or unskilled positions tends to be more localized, so that people ignore good job opportunities in other cities” (158). In light of our evidence in this paper, Moretti’s statements clearly need to be modified, at least for the case of low-skilled Chinese immigrants. Although immigrants are likely to enjoy a higher salary, this decisive shift in geographic location of employment presents some challenges for immigrants and their families. First, by definition, immigrants in these new destinations do not have easy access to immigrant organizations when they encounter practical difficulties (e.g., driving license tests, health care access, and church services in their native language). This is particularly challenging for recently arrived immigrants who do not speak English very well. Second, these Chinese immigrants are far away from their immigrant networks (often in NYC) that provide friendship or moral support and work long hours in non-gateway destinations (typically 12 hours a day for six days a week). Under such circumstances, immigrants are very vulnerable to mental health issues that deserve our attention (Luo 2006). Third, immigration scholars also study potential racial conflicts between local residents and immigrants in new destinations. Our study suggests that Chinese immigrants and entrepreneurs are generally well received and some Chinese restaurant owners also hire local white workers. For example, Mr. Lin, who used to own two restaurants in South Carolina, mentioned hiring white workers to take care of the bar in his restaurants. He said he had to pay a higher salary to white workers, but insisted that it helped his business. In our ongoing work in Northern Philadelphia, the situation is less rosy. These are neighborhoods with a high concentration of minority members, a high rate of unemployment and poverty, and a high rate of single-parent families. As such, running a restaurant in that kind of environment is often dangerous. In the first week of August 2016, 12 Chinese restaurants were robbed at gunpoint (Overseas Chinese Daily 2016). Most restaurant owners in the area have installed bullet-proof windows and security cameras. Perhaps this is not news for immigration scholars; however, the potential race and ethnic conflict needs more attention (Lee 2006; Min 1996). Finally, based on our fieldwork, we find that some immigrant workers left family members in NYC while they moved from job to job across states. This clearly generates additional stress for immigrant families, especially those with children. Thus, moving to non-gateway destinations may prove to be a mixed blessing for some families. To the extent that the immigration labor market is quickly expanding to non-traditional destinations, scholars and policymakers need to find ways to facilitate immigrant adaptation in these destinations and resolve potential issues related to these new types of fragile immigrant families. Footnotes 1 Singer (2004) calls some of the new destination cities “new immigrant gateways” (5). 2 Recent reports suggest that some rotation credit practices have gone bad and immigrants have actually suffered big financial losses in the Chinese community in NYC (World Journal 2012). 3 Massey (2008) mentioned this in passing (345) for the case of Mexican immigrants but did not elaborate. 4 We note that other researchers hold a different view of the enclave economy (Zhou 1992). 5 See https://ucr.fbi.gov/crime-in-the-u.s/2007. 6 The website location is http://downloads2.esri.com/support/whitepapers/ao_/J9509_ESRI_DataandMaps2006.pdf. 7 We estimated identical models using the 2010 survey data as well and obtained very similar results. 8 Economic historian Susan Carter (2012) documented historical evidence of Chinese immigrant settlement in non-gateway destinations. Also see Chen (2015) for a systematic treatment of historical development of Chinese restaurants in the United States. Although historically there were some Chinese restaurants in small towns and faraway places in the country, today’s dramatic expansion of Chinese restaurants is on a different scale (Lee 2009). Lee (2009) estimated that there were some 40,000 Chinese restaurants in the United States, more than the number of McDonalds, Burger Kings, and KFCs combined. Of course, the expansion of Chinese restaurants is a global phenomenon, as rising numbers of emigrants from China reach different parts of the world (Roberts 2004). About the Authors Zai Liang is Professor of Sociology at the University at Albany and Changjiang Scholar Visiting Chair Professor at Xi’an Jiaotong University. His main research interests are internal and international migration, and urban sociology. He pursues these interests in the contexts of the United States, China, and Africa. Jiejin Li is a lead Programmer/Analyst at Department of Public Health Sciences at the University of Rochester. She received her PhD in Sociology from the University at Albany. Her main research interests include migration, population health, and health care utilization. Glenn Deane is Professor and Chair of the Department of Sociology at the University at Albany. His research focus is on within-family dynamics. Zhen Li is an Assistant Professor in the Asian Demographic Research Institute at Shanghai University. Zhen’s research focuses on the patterns, changes, and consequences of internal migration in contemporary China. Bo Zhou is Ph.D. candidate in sociology at the University at Albany, SUNY. His research interests are migration, immigration, and globalization. He has participated in research projects in multiple sites in the United States and China. References Anselin , Luc . 1998 . Spatial Econometrics: Methods and Models . New York : Springer Science & Business Media . Autor , David H. 2008 . “The Economics of Labor Market Intermediation: An Analytic Framework.” NBER Working Paper 14348. Bailey , Thomas , and Roger Waldinger . 1991 . “ Primary, Secondary, and Enclave Labor Markets: A Training System Approach .” American Sociological Review 56 ( 4 ): 432 – 45 . Google Scholar Crossref Search ADS Berk , R. , and J. MacDonald . 2007 . “Overdispersion and Poisson Regression.” Unpublished paper. Bonacich , E. , and Suzanne Modell . 1980 . The Economic Basis of Ethnic Solidarity: Small Business in Japanese American Community . Berkeley : University of California Press . Card , David , and Ethan G. Lewis . 2005 . “The Diffusion of Mexican Immigrants during the 1990s: Explanations and Impacts.” In Mexican Immigration to the United States , edited by George J. Borjas , 193 – 228 . Chicago : University of Chicago Press . Carter , Susan B. 2012 . “Embracing Isolation: Chinese American Geographic Redistribution during the Exclusion Era, 1882–1943.” Paper presented at the Annual Meeting of the Population Association of America, San Francisco, CA. Chen , Yong . 2015 . Chop Suey, USA: The Story of Chinese Food in America . New York : Columbia University Press . Chi , Guangqing , and Jun Zhu . 2008 . “ Spatial Regression Models for Demographic Analysis .” Population Research and Policy Review 27 : 17 – 42 . Google Scholar Crossref Search ADS Crowder , Kyle , and Scott J. South . 2008 . “ Spatial Dynamics of White Flight: The Effects of Local and Extralocal Racial Conditions on Neighborhood Out-Migration .” American Sociological Review 73 ( 5 ): 792 – 812 . Google Scholar Crossref Search ADS PubMed Dolnick , Sam . 2011 . “Many Immigrants Job Search Start in Chinatown.” New York Times, February 22. http://www.nytimes.com/2011/02/23/nyregion/23chinatown.html, accessed March 18, 2017. Donato , Katherine M. , Charles Tolbert , Alfred Nucci , and Yukio Kawanno . 2008 . “Changing Faces, Changing Places: The Emergence of New Non-Metropolitan Immigrant Gateways.” In New Faces in New Places: The Changing Geography of American Immigration , edited by Douglas S. Massey , 75 – 98 . New York : Russell Sage Foundation . Fennelly , Katherine . 2005 . “Latinos, Africans, and Asians in the North Star State: Immigrant Communities in Minnesota.” In Beyond Gateway: Immigrants in A Changing America, edited by Elzbieta M. Gozdziak and Susan F. Martin , 111 – 36 . Lanham, MD : Lexington Books . Flippen , Chenoa A. , and Eunbi Kim . 2015 . “ Immigrant Context and Opportunity: New Destinations and Socioeconomic Attainment among Asians in the United States .” Annals of the American Academy of Political and Social Science 660 : 175 – 98 . Google Scholar Crossref Search ADS Foner , Nancy . 2000 . From Ellis Island to JFK: New York’s Two Great Waves of Immigration . New Haven, CT : Yale University Press . Fong , Eric , Chiu Luk , and E. Ooka . 2005 . “ Spatial Distribution of Suburban Ethnic Businesses.” Social Science Research 34 : 215 – 35 . Google Scholar Crossref Search ADS Gozdziak , Elzbieta M. , and Susan Martin . 2005 . Beyond Gateway: Immigrants in a Changing America . New York : Lexington Books . Granovetter , Mark . 1995 . “The Economic Sociology of Firms and Entrepreneurs.” In The Economic Sociology of Immigration: Essays in Networks, Ethnicity and Entrepreneurship , edited by Alejandro Portes , 128 – 65 . New York : Russell Sage Foundation . Greene , W. H. 2003 . Econometric Analysis , 5th ed . New York : Prentice Hall . Hilgers , Lauren . 2014 . “The Kitchen Networks: America’s Underground Chinese Restaurant Workers.” New Yorker, October 14. http://www.newyorker.com/magazine/2014/10/13/cooka%C2%80%C2%99s-tale, accessed March 18, 2017. Hirschman , Charles , and Douglas S. Massey . 2008 . “Places and People: The New American Mosaic.” In New Faces in New Places: The Changing Geography of American Immigration, edited by Douglas S. Massey , 1 – 21 . New York : Russell Sage Foundation . Kasinitz , Philip , and Jan Rosenberg . 1996 . “ Missing the Connection: Social Isolation and Employment on the Brooklyn Waterfront .” Social Problems 43 : 180 – 96 . Google Scholar Crossref Search ADS Kwong , Peter . 1997 . The Forbidden Workers: Illegal Chinese Immigrant and American Labor . New York : New Press . Leach , Mark A. , and Frank D. Bean . 2008 . “The Structure and Dynamics of Mexican Migration to New Destinations in the United States.” In New Faces in New Places: The Changing Geography of American Immigration, edited by Douglas S. Massey , 51 – 74 . New York : Russell Sage Foundation . Lee , Jennifer . 2006 . Civility in the City: Blacks, Jews, and Koreans in Urban America . Cambridge, MA : Harvard University Press . Lee , Jennifer . 2009 . Fortunate Cookies Chronicles: The Adventures in the World of Chinese Food . New York : Hachette Book Group . Li , Wei . 1998 . “ Anatomy of a New Ethnic Settlement: The Chinese Ethnoburb in Los Angeles .” Urban Studies 35 : 479 – 502 . Google Scholar Crossref Search ADS Liang , Zai , Miao David Chunyu , Guotu Zhuang , and Wenzhen Ye . 2008 . “ Cumulative Causation, Market Transition, and Emigration from China .” American Journal of Sociology 14 : 706 – 37 . Google Scholar Crossref Search ADS Lichter , Daniel , Domerico Parisi , Michael C. Taquino , and Steven Michael Grice . 2010 . “ Residential Segregation in New Hispanic Destinations: Cities, Suburbs, and Rural Communities Compared .” Social Science Research 39 : 215 – 30 . Google Scholar Crossref Search ADS Light , Ivan . 2006 . Deflecting Immigration: Networks, Markets, and Regulation in Los Angeles . New York : Russell Sage Foundation . Light , Ivan , and Edna Bonacich . 1988 . Immigrant Entrepreneurs: Koreans in Los Angeles . Berkeley : University of California Press . Light , Ivan , and Carolyn Rosenstein . 2005 . “Expanding the Interaction Theory of Entrepreneurship.” In The Economic Sociology of Immigration: Essays on Networks, Ethnicity, and Entrepreneurship , edited by Alejandro Portes , 166 – 212 . New York : Russell Sage Foundation . Light , Ivan , George Sabagh , Mehdi Bozorgmehr, and Claudia Der-Martirosian . 1994 . “ Beyond the Ethnic Enclave Economy .” Social Problems 41 : 65 – 80 . Google Scholar Crossref Search ADS Logan , John R. , Richard Alba , and Thomas L. McNulty . 1994 . “ Ethnic Economies in Metropolitan Regions: Miami and Beyond .” Social Forces 72 : 691 – 724 . Google Scholar Crossref Search ADS Logan , John R. , Richard Alba , and Wenquan Zhang . 2002 . “ Immigrant Enclaves and Ethnic Communities in New York and Los Angeles .” American Sociological Review 67 : 299 – 322 . Google Scholar Crossref Search ADS Long , Scott . 1997 . Regression Models for Categorical and Limited Dependent Variables . Thousand Oaks, CA : Sage Publications . Long , J. Scott , and Jeremy Freese . 2006 . Regression Models for Categorical Dependent Variables . Using Stata, 2nd ed . College Station, Texas : Stata Press . Luo , Michael . 2006 . “Immigrants Hear God’s Word, in Chinese, by Conference Call.” New York Times, May 21. http://query.nytimes.com/gst/fullpage.html?res=9C03E1D6133EF932A15756C0A9609C8B63&sec=&spon=&pagewanted=2, accessed February 26, 2009. Marrow , H . 2011 . New Destinations Dreaming: Immigration, Race, and Legal Status in the Rural American South . Stanford, CA : Stanford University Press . Massey , Douglas S . 1995 . “ The New Immigration and the Meaning of Ethnicity in the United States .” Population and Development Review 21 : 631 – 52 . Google Scholar Crossref Search ADS Massey , Douglas S. , ed. 2008 . New Faces in New Places . New York : Russell Sage Foundation . Massey , Douglas S. , and Chiara Capoferro . 2008 . “The Geographic Diversification of American Immigration.” In New Faces in New Places: The Changing Geography of American Immigration , edited by Douglas S. Massey , 343 – 54 . New York : Russell Sage Foundation . Massey , Douglas S. , J. Durand , and K. A. Pren . 2016 . “ Why Border Enforcement Backfired.” American Journal of Sociology 121 : 1557 – 1600 . Google Scholar Crossref Search ADS Min , Pyong Gap . 1996 . Caught in the Middle: Korean Merchants in America’s Multiethnic Cities . Berkeley, CA : University of California Press . Min , Pyong Gap . 2005 . “Settlement Patterns and Diversity.” In Asian Americans: Contemporary Trends and Issues, edited by Pyong Gap Min , 32 – 53 . 2nd ed . Thousand Oaks, CA : Pine Forge Press . Moretti , Enrico . 2012 . The New Geography of Jobs . New York : Mariner Books . Mouw , Ted . 2002 . “ Are Black Workers Missing the Connection? The Effect of Spatial Distance and Employee Referrals on Interfirm Racial Segregation .” Demography 39 ( 3 ): 507 – 28 . Google Scholar Crossref Search ADS PubMed Nee , Victor , and Brett de Bary Nee . 1973 . Longtime Californ’: Documentary Study of an American Chinatown . New York : Pantheon Books . Nee , Victor , Jimmy M. Sanders , and Scott Sernau . 1994 . “ Job Transitions in an Immigrant Metropolis: Ethnic Boundaries and the Mixed Economy .” American Sociological Review 59 : 840 – 72 . Google Scholar Crossref Search ADS Newman , Barry . 2005 . “ On the East Coast, Chinese Buses Give Greyhound a Run .” Wall Street Journal , January 25, A1 . New York City Department of City Planning . 2009 . Chinatown Bus Study . New York : New York City Department of City Planning . Office of Immigration Statistics . 2008 . 2007 Yearbook of Immigration Statistics . Washington, DC : US Department of Homeland Security . Office of Immigration Statistics . 2016 . 2015 Yearbook of Immigration Statistics . Washington, DC : US Department of Homeland Security . Overseas Chinese Daily . 2016 . “12 Chinese Restaurants in Philadelphia Were Robbed at Gun Point.” August 9. http://forums.huaren4us.com/archiver/showtopic.aspx?topicid=2051423, accessed March 18, 2017. Parrado , Emile , and William Kandel . 2008 . “New Hispanic Migrant Destinations: A Tale of Two Industries.” In New Faces in New Places , edited by Douglas S. Massey , 99 – 123 . New York : Russell Sage Foundation . Portes , Alejandro , ed. 1995 . The Economic Sociology of Immigration: Essays on Networks, Ethnicity, and Entrepreneurship . New York : Russell Sage Foundation . Portes , Alejandro , and Robert Bach . 1985 . Latin Journey . Berkeley : University of California Press . Portes , A. , and Steven Shafer . 2004 . “ Revisiting The Enclave Hypothesis: Miami Twenty-Five Years Later .” Sociology of Entrepreneurship 25 : 157 – 90 . Google Scholar Crossref Search ADS Portes, Alejandro , and Ruben Rumbaut. 2006 . Immigrant America: A Portrait . Berkeley : University of California Press . Portes , Alejandro , and Julia Sensenbrenner . 1993 . “ Embeddedness and Immigration: Notes on the Social Determinants of Economic Action .” American Journal of Sociology 98 ( 6 ): 1320 – 50 . Google Scholar Crossref Search ADS Raudenbush , Stephen W. , and Anthony S. Bryk . 2002 . Hierarchical Linear Models: Applications and Data Analysis Methods . Thousand Oaks, CA : Sage Publications . Roberts , J. A. G. 2004 . China to Chinatown: Chinese Food in the West . Chicago : University of Chicago Press . Sanders , Jimmy , and Victor Nee . 1987 . “ Limits of Ethnic Solidarity in the Enclave Economy .” American Sociological Review 52 : 745 – 67 . Google Scholar Crossref Search ADS Sassen , Saskia . 1995 . “Immigration and Local Labor Markets.” In The Economic Sociology of Immigration: Essays on Networks, Ethnicity, and Entrepreneurship , edited by Alejandro Portes , 87 – 127 . New York : Russell Sage Foundation . Scott , A. J. 1988 . Metropolis from the Division of Labor to Urban Form . Berkeley : University of California Press . Singer , Audrey . 2004 . “The Rise of New Immigrant Gateways.” The Living Cities Census Series, February . Washington, DC : Brookings Institution . Stovel , Katherine , and Lynette Shaw . 2012 . “ Brokerage .” Annual Review of Sociology 38 : 139 – 58 . Google Scholar Crossref Search ADS Waters , Mary , and Tomas R. Jimenez . 2005 . “ Assessing Immigrant Assimilation: New Empirical and Theoretical Challenges .” Annual Review of Sociology 31 : 105 – 25 . Google Scholar Crossref Search ADS Williamson , Oliver . 1975 . Markets and Hierarchy: Analysis and Antitrust Implications . New York : Free Press . Wilson , William Julius . 1987 . The Truly Disadvantaged: The Inner City, the Underclass and Public Policy . Chicago : University of Chicago Press . Wilson , Kenneth , and Alejandro Portes . 1980 . “Immigrant Enclaves: An Analysis of the Labor Market Experiences of Cuban in Miami.” American Journal of Sociology 86 : 295 – 319 . Xie , Yu , and Margret Gough . 2011 . “ Ethnic Enclaves and the Earnings of Immigrants .” Demography 48 : 1293 – 1315 . Google Scholar Crossref Search ADS PubMed Zeng , Huiyan . 2009 . “ Difficult to Find Jobs in Chinese Community in Times of Financial Crisis .” World Journal Magazine , February 22, 16 – 20 . Zhou , Min . 1992 . Chinatown: The Socioeconomic Potential of an Urban Enclave . Philadelphia : Temple University Press . Zhou , Yu . 1998 . “ Beyond Ethnic Enclaves: Location Strategies of Chinese Producer Service Firms in Los Angeles .” Economic Geography 74 : 228 – 51 . Google Scholar Crossref Search ADS Zuniga , Victor , and Ruben Hernandez-Leon . 2005 . New Destinations: Mexican Immigration in the United States . New York : Russell Sage Foundation . Author notes This project is supported by the Russell Sage Foundation (#88-10-06), whose support is gratefully acknowledged. The senior author was a visiting scholar at the Russell Sage Foundation while working on the initial version of this paper. Earlier versions of this paper were presented at Princeton University, Queens College, CUNY Graduate Center, Brown University, Sun Yat-sen University, the University of Texas at Austin, and the Russell Sage Foundation. We thank Hui-shien Tsao, Jin Lee for technical assistance, and Katie Winograd, Christine Walsh for editorial assistance. Douglas S. Massey provided important advice at the initial stage of the project, and Steve F. Messner gave important advice about access to crime data. Qian Jasmine Song, Simon Chen, and Bo Zhou contributed to our data collection in New York City. Bo Zhou created map 1 for the paper. We also acknowledge the institutional support from the Center for Social and Demographic Analysis at the University at Albany (with core funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development [2 R24 HD044943-09]). © The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - From Chinatown to Every Town: New Patterns of Employment for Low-Skilled Chinese Immigrants in the United States
JO - Social Forces
DO - 10.1093/sf/soy061
DA - 2018-12-01
UR - https://www.deepdyve.com/lp/oxford-university-press/from-chinatown-to-every-town-new-patterns-of-employment-for-low-Teg0yQPCJM
SP - 893
VL - 97
IS - 2
DP - DeepDyve
ER -