Chinese Immigrant Entrepreneurship in the United States: Temporal and Spatial Dimensions

Abstract

Focusing on transnational entrepreneurship and immigrant businesses in new destinations, this paper studies entrepreneurship of Chinese immigrants in the U.S. using data from three surveys. In the analysis of transnational connections, we focus on the temporal dimension that links pre-migration and post-migration business activities. Results from logistic models reveal that the prospect of being self-employed among Chinese immigrants is significantly enhanced if they are from households in China with business backgrounds. This finding highlights the fact that transnational entrepreneurship is embedded in the multi-stranded connections between the immigrant sending and receiving societies. In the second part of the paper, sequence analysis is used to describe and classify business trajectories in traditional and new immigrant destinations. The results establish that while it may take a longer time for immigrants to achieve business ownership in new destinations than in traditional destinations, new immigrant destinations increase the chance of business expansion from one business to multiple businesses. These findings indicate a transition in immigrant entrepreneurs’ business models. Businesses in traditional destinations mainly follow a survival strategy, while those in new destinations are adopting models that are akin to mainstream business operations, which gives rise to more opportunities for socioeconomic mobility.

Entrepreneurship is a mark of economic advancement for immigrants. As a pathway for self-sufficiency and wealth accumulation, entrepreneurship provides immigrants an effective vehicle to achieve upward mobility in the host society (Wilson and Martin 1982). Immigrant businesses also constitute an integral part of the ethnic communities they are associated with. Except for offering ethnic products and services, these businesses also provide employment opportunities to their co-ethnics (J. E. Smith 1985; Waldinger, Aldrich, and Ward 1990; Wilson and Portes 1980). Because of its importance to individual immigrants and communities in both the immigrant-receiving and sending societies, immigrant entrepreneurship has attracted scholarly attention around the world (Kariv et al. 2009; Zapata-Barrero and Rezaei 2020; Brzozowski and Cucculelli 2020). However, in the current literature, there is a lack of discussion on how temporal and spatial factors intersect to affect immigrant entrepreneurship. In this paper, we try to fill this gap by analyzing entrepreneurial behaviors of Chinese immigrants in the U.S. from a perspective that synthesizes both the temporal dimension and the spatial dimension. This new perspective sheds new light on two important topics in immigrant entrepreneurship research: transnationalism and immigrant businesses in new destinations.

Transnationalism has been established as a key feature of immigrant entrepreneurs (Portes, Haller, and Guarnizo 2002). In prior studies, transnational entrepreneurs mainly refer to those who participate simultaneously in business activities in more than one nation-state. This definition focuses on the spatial dimension of cross-border connections, but it largely overlooks the temporal dimension of transnationalism. Recent research has shown that immigrants’ pre-migration experiences have long-lasting impacts on their post-migration behaviors (Waldinger 2015). As for research on immigrant entrepreneurship, although previous research indicates that pre-migration entrepreneurial experiences and household business background can impart immigrants business aspiration and skills, financial support, and access to market information (Portes and Rumbaut 2014), the linkage between pre-migration and post-migration business activities has not been fully incorporated into the discussion on transnational entrepreneurship.

In addition to transnationalism, extant research on immigrant entrepreneurship also emphasizes its presence in ethnic enclaves. Given that the geographic distribution of immigrants becomes more diversified in recent decades (Massey 2008), the literature on immigrant entrepreneurship must also reflect this changing reality in the spatial dimension and integrate it into the careers of these business owners. As articulated by recent research on the mixed embeddedness of immigrant entrepreneurship (Kloosterman and Rath 2018; Zubair and Brzozowski 2018), the constraints and opportunities that immigrant entrepreneurs face in new destinations can be vastly different from what their predecessors are accustomed to in traditional destinations. To explore how immigrant entrepreneurs adapt to the context of geographic diversification, comparisons between traditional immigrant gateways and new destinations are clearly warranted.

In this paper, we aim to advance the literature on immigrant entrepreneurship by studying Chinese immigrants in the U.S. using data from three surveys, one in China, two in the U.S. The analysis proceeds in two steps. First, we analyze the correlation between transnational connections and entrepreneurship using our sample of Chinese immigrants. Specifically, we expand the concept of transnationalism along the temporal dimension and investigates how pre-migration exposure to business activities may promote post-migration entrepreneurship. After analyzing the correlation between pre-migration and post-migration business activities, we move on to describe and analyze the entrepreneurship trajectories of the immigrants using sequence analysis methods. The survey data used in our analysis cover immigrant business owners in both traditional and new immigrant destinations, making it possible for us to make comparisons between the two contexts. Also, the sequence analysis approach allows us to conduct a nuanced investigation on entrepreneurial careers from the temporal dimension, which helps to distinguish immigrant entrepreneurs’ socioeconomic mobility (e.g. business expansion) from the surviving strategy of living on a single business.

Background and Hypotheses

Entrepreneurship of Chinese migrants

Chinese migrants have a noticeable presence in business activities in North America, Europe, and Northeastern Asia (Kariv et al. 2009; Falavigna, Ippoliti, and Manello 2019; Ren and Liu 2015). For early Chinese immigrants in the U.S., partly because of the anti-Chinese violence and legislation, their entrepreneurial activities were confined to lines of trade that served their co-ethnics and required long hours, hard work, cheap labor, and low profit (Waldinger, Aldrich, and Ward 1990). By that time, laundry was the single largest number of businesses (Wong 1982). After 1900, the invention of chop suey helped attract white customers who were lured by the prospect of a bland, mildly exotic dish served at a cheap price (Chen 2014). Consequently, the restaurant trade started to burgeon and expand outside Chinatowns. For these immigrant entrepreneurs, diasporic organizations in Chinatowns and ethnoburbs are crucial to their business success (Zhou and Liu 2016; Li 2009): on the one hand, immigrants who engage in routine and sustained transnational activities can do so via these diasporic organizations; on the other hand, information on entrepreneurial opportunities in the U.S. is also shared within diasporic communities and co-ethnic networks. Since the turn of the 21^st century, as an increasing number of Chinese immigrants settle down and establish businesses in new immigrant destinations (Liang and Zhou 2018), their business practices and entrepreneurial careers may also take a new form.

Transnationalism and Immigrant Entrepreneurship

Transnational entrepreneurship is an effective path to immigrants’ economic adaptation (Portes, Haller, and Guarnizo 2002). Transnationalism refers to processes by which immigrants forge and sustain multi-stranded social relations that link together their societies of origin and settlement (Basch, Schiller, and Szanton Blanc 1994). While the origins of transnationalism can be traced back to transatlantic trade practiced by European immigrants in the U.S. during the turn of the 20^th century (Foner 1997), in contemporary societies, due to technological advancement and changes in immigrant sending and receiving societies, the impact of transnationalism is no longer restricted to international trade, but now spans multiple life domains, including economics, politics, culture, and religion (Levitt and Jaworsky 2007; R. Smith 2006).

Because of discrimination, language barriers, devaluation of foreign credentials, and other disadvantages in the mainstream labor market, some immigrants enter entrepreneurship to survive in the host society and achieve upward mobility (Yoon 1997; Min 1996). For these immigrant entrepreneurs, their cross-country social networks may facilitate entrepreneurial activities in the host society, including the transfer of knowledge and technology, financial flows, research and development activities, and job outsourcing (Brzozowski, Cucculelli, and Surdej 2017). Although the scope of transnational entrepreneurship has been expanded from international trade to a wider range of economic and social domains, the current literature still focuses on a spatial dimension that emphasizes simultaneous engagement in the sending and receiving societies. In this paper, we expand the concept of transnationalism along the temporal dimension and look at the linkage between immigrants’ post-migration entrepreneurial activities and their pre-migration business backgrounds.

This temporal dimension is crucial to Chinese and other Asian immigrant groups in the U.S. Unlike Hispanic immigrants who can relatively easily travel across the border, Asian immigrants face higher costs and more barriers when engaging in transpacific trade or other business activities that require frequent and sustained connections between their homeland and the U.S. Although there are some successful examples of Asian immigrants actively using their transnational network in international trade, like Koreans in the wig industry (Chin, Yoon, and Smith 1996), the geographical distance, time difference, and transportation cost, in general, can inhibit transnational entrepreneurship as defined by the spatial perspective. However, from the temporal perspective, immigrants’ pre-migration backgrounds may be an important factor that can affect their post-migration entrepreneurial activities. In the general literature on entrepreneurship (Aldrich 2005), business experiences have been established as an important source of entrepreneurial aspiration, business skills, and financial savings for future business-founding activities. For immigrants, because pre-migration experiences of self-employment are rare among most immigrant groups, the linkage between pre-migration business experiences and post-migration entrepreneurship has only been quantitatively tested in a few empirical studies (Brzozowski and Cucculelli 2020; Portes and Rumbaut 2014). In this study, with a relatively large sample, we are able to test this important temporal dimension of transnational entrepreneurship among Chinese immigrants.

It is noteworthy that immigrants’ pre-migration exposure to business activities may not wholly hinge on their personal experiences. They may also benefit from household businesses even if they have no self-employment experience before emigration. Although the prevalence of entrepreneurship among Chinese immigrants may have fostered an entrepreneurial culture in ethnically concentrated areas, immigrants from entrepreneurial households may still have some additional advantages in entering entrepreneurship. Using business profits, these households can provide financial support for the migration trip and business establishment in the host society. Moreover, the environments of these business households are conducive to the cultivation of entrepreneurial aspirations and skills. Empirical research has established that household background plays an important role in shaping people’s occupational orientation (Halaby 2003). For immigrants growing up in business households, the disposition towards entrepreneurship is deeply cultivated in their early years. Similarly, they can learn business skills by either formally or informally working in their household businesses. After they settle down in the host society, these skills and knowledge can give them advantages in organizing and operating businesses.

The above discussion on transnational business households leads to our first hypothesis.

Hypothesis 1 (temporal hypothesis): Chinese immigrants with pre-migration entrepreneurial experiences and those from entrepreneurial households are more likely to be self-employed after migrating to the U.S., as compared to those with no prior business exposure.

From Traditional to New Destinations

Our discussion on transnationalism from the temporal dimension mainly focuses on immigrants’ background in the sending society. At the receiving end of migration, there is an emerging phenomenon that immigrants are settling down in new immigrant destinations. According to the 2000 census, nearly one-third of immigrants resided outside established settlement states (Singer 2004). Empirical studies have documented this trend of moving to new destinations among various Hispanic and Asian immigrant groups (Flippen and Kim 2015; Gozdziak and Martin 2005; Marrow 2011; Massey 2008; Zúñiga and Hernández-León 2005).

This new pattern of immigrant settlement poses two major challenges for immigrant assimilation (Massey 2008; Waters and Jiménez 2005): first, unlike traditional destinations, new destinations often lack co-ethnic organizations and communities that can ease the initial settlement process for new arrivals and ultimately help immigrants assimilate into the host society; second, residents in new destinations usually do not have prior encounters with immigrants, which could lead to potentially troubling race and ethnic relations.

While there is a burgeoning literature looking into immigrants settling down in new destinations and their assimilation outcomes (Ackert, Crosnoe, and Leventhal 2019; Lichter, Sanders, and Johnson 2015; Winders 2014), little research has been done to investigate differences in immigrant entrepreneurship between new and traditional destinations. As the assimilation results tend to be different between self-employed immigrants and immigrant employees (Griffith 2008), research on immigrant business owners in new destinations will not only shed new light on patterns of immigrant entrepreneurship, but also further our understanding of assimilation processes in the context of geographic diversification of immigration.

Benefiting from the development of ethnic-based transportation networks (e.g. Chinatown bus), employment agencies, and business supply chains, an increasing number of Chinese immigrant entrepreneurs can now run businesses in new destinations but keep recruiting employees and procuring ethnic ingredients from Chinatown in New York City (Liang and Zhou 2018). In these new destinations, the political-institutional, cultural, and economic contexts are vastly different from what immigrant business owners are used to in Chinatowns, ethnoburbs¹, and other traditional destinations that have a history of Chinese settlement and entrepreneurship. According to the mixed embeddedness perspective, several features of new destinations may pose unique opportunities and constraints for immigrant entrepreneurs (Solano 2020; Kloosterman and Rath 2018; Zubair and Brzozowski 2018).

In new immigrant destinations, the major obstacle for immigrant entrepreneurs is the lack of a co-ethnic population. In traditional destinations, immigrant entrepreneurs can get financial support (e.g. rotational credit associations) and consumers from their co-ethnic communities, but in new destinations, they have to rely on their own savings and cater to non-Hispanic White customers or clientele from other minoritized groups (Waldinger, Aldrich, and Ward 1990; Bates 1997). Sometimes, those who aspire for entrepreneurship have to resettle multiple times before they eventually find a good place for their businesses (Watson 1977). These barriers suggest that establishing a business in new destinations can be more time-consuming than in traditional destinations.

Hypothesis 2 (business establishment hypothesis): It is more time-consuming for immigrants to establish businesses in new immigrant destinations than in traditional destinations.

The obstacles for immigrant businesses in new destinations mainly appear in the initial stage of business establishment. Once a business is up and running, the lack of co-ethnics turns into an advantage because there is little competition from other co-ethnic businesses (Zubair and Brzozowski 2018). Furthermore, because immigrant businesses tend to fill in market niches that are distinguishably different from businesses established by the dominant or other minoritized groups, competition from other racial or ethnic groups is also likely to be at a low level (Waldinger, Aldrich, and Ward 1990). In addition to less competition, new destinations also tend to have lower commercial rental prices compared to established destinations like New York City. Together, these benefits can make it easier for immigrant businesses to expand than in traditional destinations.

Hypothesis 3 (business expansion hypothesis): Immigrant businesses in new destinations are more likely to expand from one business to multiple businesses than businesses in traditional destinations.

Data and Methods

Data

Data used in this paper are from three surveys, one in China (2000–2002) and two in the US (2004 and 2014).

The 2000–2002 China survey and 2004 US survey are both from the China International Migration Project, which follows the successful model of the Mexican Migration Project and the Latin American Migration Project (Durand and Massey 2004; Massey 1987). The China International Migration Project interviewed one sample of respondents in China from 2000 to 2002 and another sample in the U.S. in 2004. The China sample was selected from eight towns within northeastern Fujian Province that are known to send large numbers of migrants to the U.S. For each town, a target sample of 200 households was drawn from stratified sampling. The final sample size is 1,340 households. The U.S. sample comprises a nonrandom snowball sample of immigrants in New York City who are from these eight towns. For each of the eight towns, about 40 to 50 immigrants in New York City were interviewed, with the final sample size being 410.

The 2014 U.S. survey focuses exclusively on Chinese immigrant entrepreneurs in new destinations. In total, 36 Chinese immigrant entrepreneurs were selected from six states (Florida, North Carolina, Ohio, Pennsylvania, Texas, and Virginia), with six entrepreneurs in each state. Two methods of data collection were used to select these 36 entrepreneurs. The first method relied on religious networks to identify Chinese immigrant entrepreneurs in the six states. Because a very high proportion of Chinese immigrants (especially immigrants from Fujian province) are religious, using religious connections as a way to identify immigrant entrepreneurs is an effective way to carry out the survey. The second way to identify Chinese immigrant entrepreneurs (restaurant owners in most cases) is through the use of Google maps. Once a location² in one of the six states was selected, the interviewers used Google maps to list all Chinese restaurants within 50 miles of that location. Then they randomly called each restaurant on the list until the target sample size was reached. In addition to administering a standardized questionnaire, the interviewers also conducted in-depth interviews with at least two business owners in each state.

Though each of the three surveys has its own strengths and limitations, together they provide a more complete picture of entrepreneurial experiences for immigrants. As for the 2000–2002 China survey, most of the respondents were left-behind household members rather than the migrants themselves. These non-migrants are usually well informed of the status of their emigrant family members, either through direct long-distance communication or based on information provided by other acquaintances and villagers. However, they are not able to provide detailed retrospective data on the migrants’ adaptation experience in the U.S. Thus, we can only conduct cross-sectional analyses using this data source.

In contrast, the 2004 and 2014 U.S. surveys interviewed immigrants themselves, making it possible for us to conduct longitudinal analysis using retrospective data about business activities in the U.S. In this study, we pool data from the two surveys to compare entrepreneurial activities in traditional destinations and new destinations. It should be noted that the key strength of the 2004 and 2014 surveys is information on ownership and timings of multiple businesses and their spatial locations, which is not available in other national surveys. This allows us to gain important insights, which are not possible otherwise. However, because there is a time lag of about 10 years between the two surveys, we realize that our comparison between traditional and new destinations can be confounded by other factors that have changed from 2004 to 2014. For example, the 2008 global financial crisis may have led to disruptions in immigrants’ entrepreneurial careers. These impacts would be captured by the 2014 survey but not the 2004 survey.

Methods

Logistic Regression Using the 2000–2002 China survey

To test our first hypothesis, which is about the correlation between transnationalism and entrepreneurship, we need to compare immigrant business owners with immigrant employees. Because the hypothesis requires measures for household business in China, the 2000–2002 China survey is the only data source that we can use. The other two surveys mainly focus on the immigrants’ own adaptation and economic activities in the U.S., with very limited information about their household background in China.

The sample from the China survey was designed to be representative of migrant-sending communities at the household level. Thus, if we estimate models with individual immigrants as the unit of analysis, we will lose the representativeness of our data because some households have sent more immigrants to the U.S. than others. To address this issue, the analytic sample used in this part is selected through a re-sampling procedure that allows each household to contribute only one individual who was born in China, was at least 15 years old but not retired, and was in the U.S. at the time of the survey. When a household has sent more than one immigrant, we randomly draw one individual from all qualified immigrants in the household.

The major limitation of the China survey is that we do not have retrospective data about the immigrants’ business activities in the U.S. Because of this limitation, we cannot use event history models to study transition into entrepreneurship from a longitudinal perspective. Instead, we estimate logistic regression models to compare business owners and employees.

The dependent variable is whether an immigrant in the U.S. was self-employed at the time of the survey, coded as 1 if yes and 0 otherwise. This information is drawn from the immigrant’s current occupation in the U.S.

The immigrants’ business background is measured with three variables. At the individual level, a dummy variable measuring immigrants’ own pre-migration business experience is constructed from their job histories in China. At the household level, two dummy variables are used to measure household business background in China. Both of them are derived from the household business history module in the questionnaire, which provides the starting and closing years of all household businesses. The first household-level variable measures whether the household had ever been engaged in business activities before any household member went abroad. The second one measures whether the household owned active business in China at the time of the survey. These two household-level measures are complementary to each other. The first one can be used to test the hypothesis about transnationalism from the temporal perspective. The strict temporal order between the independent and dependent variables also allows us to make a causal inference. The second variable can be used to capture any possible simultaneous and substantive linkage between businesses in the two countries.

Some other personal and household characteristics are also included in the analysis as control variables. At the individual level, we control for the immigrant’s sex, age, marital status, education, religious affiliation, legal status, number of years in the U.S., and cost of migration. Control variables at the household level include the number of household members in the U.S. and the total migration cost of household members other than the focal migrant. Descriptive statics of variables used in the logistic models can be found in Table 1.

Table 1:

Descriptive statistics on Chinese immigrants by self-employment status at the time of the survey, the 2000–2002 China survey

Variables	Self-Employed	Not Self-Employed	Combined
Sex
Male	60.99	72.55	71.22
Female	39.01	27.45	28.78
Age
19 or younger	0.00	3.14	2.78
20–24	4.96	19.78	18.07
25–29	21.99	22.92	22.81
30–34	30.50	18.21	19.62
35–39	16.31	14.70	14.88
40–44	12.06	9.89	10.14
45–49	10.64	6.38	6.87
50 or older	3.55	4.99	4.82
Marital status
Ever married	95.04	64.88	68.36
Never married	4.96	35.12	31.64
Education
Elementary school or less	26.95	27.17	27.15
Junior high school	46.10	54.71	53.72
Senior or vocational high school	25.53	16.36	17.42
College or above	1.42	1.76	1.72
Religious affiliation
Christianity	5.67	5.73	5.72
Other religions	42.55	46.30	45.87
None	51.77	47.97	48.41
Have a “Green Card”
Yes	53.19	18.76	22.73
No	46.81	81.24	77.27
Had pre-migration business experience
Yes	7.80	5.64	5.89
No	92.20	94.36	94.11
From household with prior business history
Yes	17.02	13.22	13.65
No	82.98	86.78	86.35
Household currently has active business in China
Yes	6.38	7.76	7.60
No	93.62	92.24	92.40
N. of years in the U.S.	10.51	6.49	6.95
N. of household members in the U.S.	3.25	2.00	2.14
Own migration cost (2007 $)	10016.79	28928.25	26747.94
Migration cost for other household members (2007 $)	31382.16	20176.56	21468.46
N.	141	1082	1223

Note: Mean for continuous variables and percentage for categorical variables.

Sequence Analysis Using the 2004 and 2014 U.S. surveys

After comparing immigrant entrepreneurs and employees with logistic regression models, the second part of our study explores how immigrant entrepreneurs fare over time by classifying, describing, and comparing their business trajectories with sequence analysis. This analysis requires longitudinal data. Thus, we use pooled data from entrepreneurs in the 2004 and 2014 U.S. surveys. Both of these two surveys provide the full business history of each immigrant, making it possible for us to study the transition into entrepreneurship as well as the expansion from one business to multiple businesses. Because our focus here is on immigrant business owners, respondents who were covered in the 2004 survey but had never owned businesses are excluded from the analysis. Also, respondents in these two surveys are all immigrants in the U.S. Thus, there is no need for us to resample immigrants from their households. Instead, we use household heads as our analytic sample.

The sequence analysis proceeds in three steps: constructing sequences, calculating dissimilarity scores, and classifying sequence clusters. We construct sequences of entrepreneurial careers alongside two dimensions: number of businesses and location of businesses. For the first dimension, we consider the establishment of the first business and the transition from one business to two or more businesses as important signals of entrepreneurial success. Thus, this number dimension is operationalized into three categories: no business, single business, or multiple businesses. As for business locations, following Massey and Capoferro’s (2008) practice, we define traditional and new destinations at the state level. Specifically, we consider three states (California, Illinois, and New York) that have a high concentration of Chinese immigrants as traditional destinations, and all other states are counted as new destinations³. The number and location dimensions classify all person-year records into five entrepreneurial states: no business, single business in traditional destinations, multiple businesses in traditional destinations, single business in new destinations, and multiple businesses in new destinations. Using these defined entrepreneurial states, we construct a trajectory (sequence) for each respondent from the year they arrived in the U.S. to the time of the survey (2004 or 2014).

In the second step, we calculate a dissimilarity score for every pair of sequences using the optimal matching algorithm (Abbott and Tsay 2000). The dissimilarity is systematically determined by “transformative costs” between each pair of sequences, namely the minimum number of operations (replacement, insertion, and deletion) needed to transform one sequence into the other. Because the immigrants came to the U.S. in different years, the sequences constructed in the first step do not have the same length. This variation in sequence length leads to a problem that longer sequences tend to have higher dissimilarity scores than shorter sequences. To address this issue, we standardize the total transformation costs with the length of the sequences (Abbott and Tsay 2000; Brzinsky-Fay, Kohler, and Luniak 2006). These standardized costs are used as dissimilarity scores for each pair of sequences.

Using the matrix of dissimilarity scores calculated in the second step, we then proceed to reduce the number of sequences into a few substantively distinct groups through cluster analysis with Ward’s (1963) method. It forms hierarchical groups of mutually exclusive subsets by maximizing the similarity within each subset. The optimal number of clusters is evaluated using the Calinski-Harabasz cluster stopping index (Calinski and Harabasz 1974). It turns out that the 2-cluster solution (Calinski-Harabasz index = 22.67) and the 3-cluster solution (Calinski-Harabasz index = 22.08) are similar in their goodness of fit. We choose to use the 3-cluster solution because it makes a theoretically informative distinction between a single business and multiple businesses in new destinations.

Compared to event history and other analytic methods that are used in the extant literature on immigrant entrepreneurship, our sequence analysis approach has two advantages (Abbott and Tsay 2000). First, instead of estimating transition rates, sequence analysis describes whole sequences of diverse events from a holistic approach. While event history models are appropriate to study transitions, the five qualitatively distinguishable entrepreneurship states in our study create too many types of transitions for these analytic methods to handle. Sequence analysis, in contrast, is useful in this case because it embeds all transitions as well as durations between the transitions into a single sequence and can thus succinctly describe the sequences and identify major patterns. Second, sequence analysis makes no modeling assumption and is consequently suited for exploration and description of dissimilarity between sequences. According to a review by Abbott and Tsay (2000), the method can identify distinctive patterns from both small and large datasets.

Results

Comparison Between Business Owners and Employees

Table 1 provides descriptive profiles of Chinese immigrant entrepreneurs and employees. Out of the total 1,233 immigrants included in the analysis, 141 (11.53%) are self-employed. A closer look at the composition of these Chinese immigrant entrepreneurs reveals that they are predominantly (82.98%) restaurateurs. Compared to immigrant employees, self-employed Chinese immigrants are more likely to have pre-migration business backgrounds at both individual and household levels. However, the business owners are slightly less likely to have active household businesses in China than employees at the time of the survey. While these results provide indicative support to our hypothesis about transnationalism from the temporal perspective, the descriptive statistics also reveal that the self-employed and employees are different in marital status, legal status, length of residence in the US, migration cost, and household size. Thus, it is necessary for us to control for these potential confounding variables and estimate the net correlation between business background in China and self-employment in the U.S.

Table 2 presents logistic regression models predicting Chinese immigrants’ self-employment status. Since the age group of 19 or younger contains no self-employed immigrants, it is combined with the 20–24 group in the models. Because of the potential collinearity between personal and household business backgrounds, the effects of these variables are estimated in separate models. Model 1 uses the immigrant’s own pre-migration experience. Model 2 uses the migrant household’s prior business background. On the basis of Model 2, Model 3 adds whether the migrant’s household currently has active business operations in China into the model.

Table 2:

Logistic regression models predicting self-employment of Chinese immigrants, the 2000–2002 China survey

	Model 1		Model 2		Model 3
Independent Variables	B	SE	B	SE	B	SE
Intercept	−6.73**	0.85	−6.69**	0.83	−6.69**	0.83
Male	−0.66**	0.24	−0.63**	0.24	−0.63**	0.24
Age
24 or younger	1.80*	0.74	1.61*	0.73	1.61*	0.73
25–29	2.09**	0.59	1.97**	0.58	1.97**	0.58
30–34	1.84**	0.56	1.73**	0.55	1.73**	0.55
35–39	1.31*	0.58	1.23*	0.57	1.23*	0.57
40–44	0.94	0.59	0.89	0.59	0.89	0.59
45–49	0.98	0.62	0.91	0.60	0.91	0.60
50 or older (reference)	----	----	----	----	----	----
Ever married	1.74**	0.48	1.70**	0.48	1.70**	0.48
Education
Elementary school or less (reference)	----	----	----	----	----	----
Junior high school	−0.09	0.25	−0.14	0.26	−0.14	0.26
Senior or vocational high school	0.56†	0.30	0.53†	0.30	0.53†	0.30
College or above	−0.78	0.90	−0.81	0.89	−0.81	0.89
Religious affiliation
Christianity	−0.20	0.45	−0.02	0.45	−0.02	0.45
Other religions	−0.02	0.21	−0.02	0.21	0.02	0.21
None (reference)	----	----	----	----	----	----
N. of years in the U.S.	0.13**	0.03	0.14**	0.03	0.14**	0.03
Have a “Green Card”	0.78**	0.22	0.76**	0.22	0.77**	0.22
N. of household members in the U.S.	0.31**	0.07	0.31**	0.07	0.31**	0.07
Had pre-migration business experience	0.73†	0.38	----	----	----	----
From household with prior business history	----	----	0.70*	0.29	0.71*	0.29
Household currently has active business in China	----	----	----	----	−0.01	0.44
Own migration cost (2007 $)	−0.02*	0.01	−0.02*	0.01	−0.02*	0.01
Migration cost for other family members (2007 $)	0.00	0.00	0.00	0.00	0.00	0.00
−2 Log Likelihood	648.88		646.49		646.49
Chi-Square	225.41**		227.80**		227.80**
N. of Immigrants	1223		1223		1223

Note

^†P < 0.10

^*P < 0.05

^**P < 0.01.

As shown in Model 1, the immigrant’s own pre-migration business experience is marginally significantly associated with post-migration self-employment. More importantly, coming from a household with prior business history clearly has a powerful positive impact on the migrant’s prospect of entrepreneurship. According to Model 2, having such a background can double (e^0.7039 = 202.16%) the odds of being self-employed in the U.S.

Model 2 suggests a positive association between pre-migration household business background and post-migration entrepreneurship, but there are two potential explanations. The traditional transnational entrepreneurship perspective, which mainly emphasizes concurrent and mutually dependent business activities in the sending and receiving societies, may suggest that the migrants are taking advantage of their ongoing ties with the households’ businesses back in China to facilitate their business activities in the U.S. The temporal perspective that we emphasize in this paper, instead, may interpret the positive impact of pre-migration business background as a reflection of financial support from households and business aspirations or skills cultivated in the immigrants’ early years.

To further adjudicate between these two interpretations, we turn to Model 3, where we simultaneously estimate the effects of pre-migration household background and the household’s current business status in China. The results suggest that migrants’ self-employment status in the U.S. is not tied to active household businesses in China. Also, in a model where we do not control for pre-migration household business (results available upon request), the coefficient of current household business in China is still non-significant. This point can be further corroborated by the comparison of specific business types between the two national settings. In the U.S., consistent with previous studies (Waldinger, Aldrich, and Ward 1990), the self-employed Chinese immigrants in our sample are mostly in the restaurant business, whereas in China, our data indicate that the migrants’ family businesses are predominantly factories and stores (34.34% and 33.48% respectively), and others mainly engage in professional or personal services (19.31%) or some other forms of middleman trade (9.87%). Among all businesses owned by the households in our sample, only 3.86% are in the restaurant business. Given such a stark difference between business activities in these two places, it is unlikely that these immigrant entrepreneurs are actively using their social ties or business linkages in China.

It is noteworthy that household business prior to migration stays significant even after the current business status in China is controlled. Taking all evidence into account, we interpret the positive effect of prior business background as an indication of possible financial support from households and the transport of entrepreneurial aspirations and skills. Although the long distance between the two countries makes it hard for immigrants to mobilize resources in their hometowns and directly benefit from their transnational networks, those from business households still have advantages in entrepreneurship no matter whether the household businesses survive after their migration.

Models in Table 2 also suggest that duration in the U.S. and permanent resident status are positively associated with self-employment. As expected, the immigrant’s own migration cost is negatively correlated with business ownership. A large share of immigrants in the survey paid fees to come to the U.S. and they must pay back the fees first before considering a new business initiative. Age appears to be curvilinearly associated with entrepreneurship, with immigrants in their late 20s most likely to be self-employed. The effects of education, religious affiliation, and migration costs of other household members are not statistically significant at the .05 level. These patterns of correlation between social-demographic characteristics and self-employment are generally consistent with studies using census data or surveys from other immigrant groups (Griffith 2008; Yoon 1997).

In terms of gender differences, although descriptive results in Table 1 shows there are more entrepreneurial men than women, models in Table 2 suggest that female immigrants have an advantage over their male peers in entrepreneurship when their sociodemographic backgrounds are held constant. This result can be explained by the fact that immigrant businesses are still dominated by the mom-and-pop business model. In international migration, female immigrants tend to migrate after their husbands or other family members have already settled down in the destination. For women whose husbands are running businesses, it is natural and logical for them to participate in family businesses. Although previous research demonstrates that only men would consider themselves as self-employed under these circumstances (Portes and Jensen 1989), respondents that we interviewed in the Philadelphia area indicated that Chinese immigrants are now registering their wives as business owners because of the bourgeoning gender equality ideology and the benefits that business ownership confers in “green card” application.

Entrepreneurship Trajectories

Figure 1 displays the three clusters of entrepreneurship trajectories identified from sequence analysis (sequence index plots). These plots show individual immigrants’ business sequences over time. The immigrant business owners are arrayed along the y-axis, with each horizontal line representing one immigrant’s entrepreneurial sequence. The x-axis is time, in our case, the number of years since the immigrant came to the U.S.

Figure 1:

Sequence index plots of entrepreneurship trajectories, pooled data from business Owners in the 2004 and 2014 US surveys

The immigrant entrepreneurs are for the most part evenly distributed across the three clusters. Immigrants in the first cluster (27 out of 74, or 36.49%) are predominantly in traditional destinations. Only four of them had established businesses in new destinations, but none of these four businesses survived longer than three years. Also, it is noteworthy that business owners in this group seldom experienced business expansion. Only one sequence shows a short spell of multiple businesses. To sum up, the first cluster features single business ownership in traditional destinations. Thus, we name it the single-traditional group.

Immigrants in the second cluster (28 out of 74, or 37.84%) also tend to hold one business only over their entrepreneurial career. Only one of them had owned multiple businesses for a brief period. However, different from those in the first cluster, immigrants in the second cluster predominantly owned businesses in new destinations. Thus, we term it as the single-new group.

The last cluster (19 out of 74, or 25.68%) features multiple businesses in new destinations. Even when the business did not expand, it tends to have been in operation for a longer time than businesses in the second group. Considering that almost all immigrants who have owned multiple businesses are in this last cluster and that these businesses are mainly in new destinations, we describe it as the multiple-new group.

In addition to the number and location of businesses, time is another dimension that is inherent in the sequence index plots. If we inspect the timing of transitions from no business to single business along the x-axis, it is quite clear that immigrant business owners in the multiple-new group have spent a longer time than those in the other two groups on preparation for the transition to self-employment. In fact, the average length of time between arriving in the U.S. and establishing the first business is 10.05 years for the multiple-new group, while for the single-traditional and single-new groups, it is 7.70 years and 6.18 years respectively (Table 3). These results suggest that it can take a longer time for immigrants to achieve entrepreneurship in new destinations than in traditional destinations. However, these time investments and efforts usually pay off. The long preparation time for business establishment also spells a promising prospect of future business expansions.

Table 3:

Descriptive statistics on Chinese immigrant entrepreneurs and their businesses in the US by entrepreneurship trajectory group, pooled data from business Owners in the 2004 and 2014 US surveys ^a

	Combined b	Single-Traditional b	Single-New b	Multiple-New b
Personal characteristics
Sample
2004	51.35	100.00	32.14	10.53
2014	48.65	0.00	67.86	89.47
Sex
Male	66.22	70.37	50.00	84.21
Female	33.78	29.63	50.00	15.79
Age at survey
25–29	15.07	7.41	32.14	0.00
30–34	12.33	7.41	21.43	5.56
35–39	13.70	14.81	17.86	5.56
40–44	16.44	7.41	10.71	38.89
45–49	17.81	22.22	14.29	16.67
50+	15.07	40.74	3.57	33.33
Marital status at survey
Ever married	93.24	96.30	89.29	94.74
Never married	6.76	3.70	10.71	5.26
Education
Elementary school or less	31.51	37.04	21.43	38.89
Junior high school	31.51	44.44	25.00	22.22
Senior or vocational high school	31.51	18.52	42.86	33.33
College or above	5.48	0.00	10.71	5.56
Religious affiliation
Christianity	16.33	7.41	33.33	14.29
Other religions	24.49	25.93	13.33	42.86
None	59.18	66.67	53.33	42.86
Age at migration
19 or younger	24.66	11.11	39.29	22.22
20–24	20.55	11.11	21.43	33.33
25–29	17.81	18.52	10.71	27.78
30–34	12.33	22.22	7.14	5.56
35–39	16.44	25.93	14.29	5.56
40–44	6.85	11.11	7.14	0.00
45–49	0.00	0.00	0.00	0.00
50 or older	1.37	0.00	0.00	5.56
Marital status at migration
Ever married	43.24	77.78	46.43	42.11
Never married	56.76	22.22	53.57	57.89
N. of years in the U.S.	15.68	15.04	11.00	23.47
N. of years in the U.S. before 1st business	7.73	7.70	6.18	10.05
N. of children at survey	2.09	2.26	1.75	2.37
N. of children at migration	0.91	1.63	0.68	0.21
Business characteristics
Ownership
Self-owned	54.46	67.74	58.06	44.00
Family-owned	20.54	6.45	25.81	26.00
Non-family partnership	13.39	25.81	12.90	6.00
Mixed partnership	11.61	0.00	3.23	24.00
N. of employees	5.84	4.10	6.24	6.68
N. of family employees	1.65	1.37	1.81	1.71
N. of business owners	74	27	28	19
N. of businesses	111	30	31	50

Notes

^a.t-tests of cross-group differences are summarized in endnote 2.

^b.Mean for continuous variables and percentage for categorical variables.

These three distinct groups of entrepreneurship trajectories indicate that immigrant business owners are by no means a homogeneous group. Because the small sample size of business owners bars us from multivariate analysis, we only further explore this heterogeneity by comparing descriptive profiles across the three groups identified in the sequence analysis (Table 3) ⁴. Here we want to highlight three preliminary findings.

First, the decision about going to new destinations versus traditional gateways is associated with the migrants’ social-demographic characteristics. Compared to immigrants with businesses in traditional destinations, those in the two new destination groups tend to be younger and less obliged to their families when migrating to the U.S. (see the age and marital status at migration). These differences indicate that people who are risk-averse (e.g. elder migrants, married people, and those who have children) are reluctant to endure the high initial risk of going to new destinations.

Second, at the business level, the three groups tend to be different in ownership structure. More than two-thirds (68.84%) of businesses in the single-traditional group are self-owned. In the single-new and multiple-new groups, the share of self-owned businesses drops to 58.06% and 44.00% respectively. Family-owned businesses, in contrast, are more prevalent in new destinations than in traditional destinations. While family-owned businesses account for only 6.45% in the single-traditional group, more than a quarter of businesses in the two new destination groups are family-owned. Also, it is noteworthy that 24% of businesses in the multiple-new group are in mixed partnership. These results suggest that businesses in new destinations tend to rely more on family members than in traditional destinations. However, as businesses expand, they also need to move away from this traditional mom-and-pop model of business operation and collaborate with non-family partners. We argue this transition from family-owned businesses to multiple partner ownership is a new model frequently observed in new destinations. This new pattern has been largely overlooked in recent research using the census or ACS data (Portes and Yiu 2013) as these data sources do not have full business history data.

Last, business expansion in new destinations spells opportunities for upward mobility. Income or business profits would be the best indicators for socioeconomic status. However, due to privacy concerns, our respondents are reluctant to give us this information. Instead, we use the number of employees per business as an indicator of business size. Comparison across the three trajectory clusters shows that businesses in the single-new and multiple-new clusters have 6.24 and 6.68 employees respectively, significantly more than those in the traditional destination cluster (4.10). Moreover, businesses in the multiple-new group also have slightly more employees than those in the single-new group. Despite the small sample size, these results are indicative that business expansion in new destinations is not reserved for small delis, but instead provides a clear pathway for upward mobility for immigrant entrepreneurs.

Discussion and Conclusions

Using data from three surveys, this paper explores new patterns of entrepreneurship of Chinese immigrants in the U.S. First, we investigated transnational entrepreneurship among these immigrants with a focus on the temporal dimension. This focus allows us to study how multi-faceted connections between the sending and receiving societies stimulate immigrant entrepreneurship more comprehensively than previous research. The results suggest that immigrants’ pre-migration business background has a stimulating effect on self-employment in the U.S. While the effect of personal pre-migration business experience is only marginally significant, the odds of being self-employed increased remarkably if the immigrants came from households that had owned businesses in China. In addition, we also found that whether the household business was still active does not provide additional explanatory power to the possibility of being an immigrant entrepreneur in the U.S. These results indicate an important distinction between Chinese and Hispanic immigrants regarding the linkage between transnationalism and entrepreneurship. While Hispanic immigrants could actively use their cross-border networks for business purposes (Portes, Haller, and Guarnizo 2002), for Chinese immigrants, the cost of this type of transnational entrepreneurship can be high. However, from the temporal perspective, those from business households still benefit from financial support from business profits and the entrepreneurial motivation and skills cultivated in their pre-migration years. We realize that the data are from the early 2000s. China has experienced transformative changes over the last two decades as its market transition deepened. Today, Private enterprises account for more than 60% of China’s GDP growth (Xinhua 2018). Thus, we expect this connection between immigrants’ household businesses in China and their entrepreneurial activities in the U.S. to be even stronger today than two decades ago.

Our second contribution is to build linkages between two strands of literature: one on new immigrant destinations and the other on immigrant entrepreneurship. The receiving society’s local context is crucial to immigrant businesses. In this paper, we presented how business trajectories differ between traditional immigrant gateways and new destinations. Our analysis indicates that while it can take a long time for immigrants to establish businesses in new destinations, these businesses are likely to expand. This finding supports our theoretical speculation that the lack of a co-ethnic community in new destinations makes it time-consuming for immigrant entrepreneurs to look for business locations and financially prepare for the first business. However, once a business gets established, there is less competition in new destinations than in traditional gateways. In addition, these new destinations tend to have low commercial rental costs. Recent developments in the ethnic-based transportation infrastructure (Chinatown buses) and supply chain are also important. They keep businesses in new destinations being connected to the co-ethnic support system in traditional destinations like New York City, making it possible for businesses in new destinations to order online from Chinatowns (Liang et al. 2018). These factors together provide conditions that are conducive to expansion into multiple businesses. Previous research on immigrant entrepreneurship has been focusing on financial outcomes or the business owners’ subjective assessment of their businesses (Brzozowski and Cucculelli 2020), largely overlooking business expansion as a measure of entrepreneurial performance. Our results indicate that the transition from one to multiple businesses spells a change in the model of immigrant entrepreneurship. Running a single business, in many ways, represents a survival strategy for immigrants to make a living in the new country, especially for new arrivals whose English is poor. Immigrant-owned multiple businesses, in contrast, is akin to the franchise business model, such as Subway, KFC, etc., in the mainstream economy. In the case of immigrant businesses, this franchise model is often informal, but our exploratory analysis still indicates that these businesses tend to have business partners other than family members (Table 3). While the involvement in multiple businesses means more coordination and responsibilities than the traditional mom-and-pop model, it also indicates a new path for socioeconomic mobility.

On methodological grounds, our study introduces a new method, sequence analysis, into immigrant entrepreneurship research. In addition to identifying the new pattern of multiple businesses that represents the transition from survival strategy to business investment and immigrant entrepreneurs’ assimilation into the mainstream business operation model, studying business trajectories also reveals that not all businesses succeed. The sequence index plots clearly show that many businesses actually failed in the first few months or years. These results reveal the importance of studying entrepreneurial trajectories and the potential contributions that sequence analysis can bring to the field.

Admittedly, our study has several limitations. First, because the 2000–2002 China survey does not have data on business history in the U.S., we only compared immigrant entrepreneurs and employees with cross-sectional analysis. Second, although we combined samples from two surveys to study business trajectories, the pooled sample size of immigrant business owners is still small and only allows us to conduct exploratory descriptions rather than estimating analytic models. Third, the 2014 survey has the problem of sampling on the dependent variable because it only covered those who had active businesses at the time of the survey. Lastly, our sample mainly represents Fujianese and low-skill Chinese immigrants, making our specific findings not generalizable to the increasing number of high-skill immigrants who mainly arrived in the U.S. in the past three decades. However, our theoretical discussions with respect to spatial and temporal dimensions should be relevant to any future study of entrepreneurship for high-skilled immigrants.

Despite these limitations, our findings still shed new light on transnationalism and the geographic diversification of immigration. Both of these two fields deserve more scholarly attention in future research. As for transnationalism, with the availability of new technology of communication, potential business connections between origin and destination societies should be stronger in the future. Also, cross-border connections can go beyond economic activities and have cultural or even political implications in the two societies (Waldinger 2015). Our research shows that cross-border connections embedded in transnational entrepreneurship can last even longer than the businesses themselves and these linkages can be sustained through kinship ties. Future studies can further explore transnational connections in cultural, political, and other domains of social lives.

Finally, students of immigrant entrepreneurship should consider entrepreneurship in new destinations in a more systematic way and test theories for multiple groups. When immigrant entrepreneurs get into new destinations, they may interact with the local residents in various ways depending on their own backgrounds and the local context (Flippen and Kim 2015; Grey and Woodrick 2005; Hernández-León and Zúñiga 2003). While our results show that conditions in new destinations can stimulate Chinese-owned businesses (mainly restaurants) to expand, patterns may be different for other immigrant groups. Also, an increasing presence of immigrant businesses may also affect other minority groups in new destinations (Somashekhar 2019). Do these businesses mainly compete with local businesses or bring job opportunities to the neighborhoods (see Table A1 in Appendix)? These issues should be further explored in follow-up studies.

Table A1:

County-level population profile of Chinese-own businesses, the 2014 US Survey

	N	% white	% black	% Hispanic	% Chinese	Average family income	% High school graduate
Florida	6	60.63	12.47	23.35	0.34	249756.30	70.76
North Carolina	6	53.96	27.65	11.33	0.60	241270.10	67.23
Ohio	6	72.23	20.68	2.44	0.40	383237.10	69.50
Pennsylvania	6	41.81	38.19	11.37	1.93	408967.00	66.82
Texas	6	49.39	7.10	40.27	0.24	409456.20	62.10
Virginia	6	60.28	22.58	7.23	0.48	388849.60	71.99
Combined	36	56.38	21.44	16.00	0.66	346922.70	68.07

Note: Data from the American Community Survey, 2010–2014 5-year average (Ruggles et al. 2020). When an immigrant entrepreneur owns multiple businesses at the time of the survey, the major business identified by the entrepreneur is used.