Social networks for cross-border business activities: a comparison between transnational and domestic Moroccan migrant entrepreneurs

ABSTRACT

The existing literature shows that social contacts and resources available through social networks are particularly important for migrant entrepreneurs. This seems to apply also to transnational migrant entrepreneurs, who are migrants involved in entrepreneurial activities crossing national borders. It is still not entirely clear which kind of social contacts they use, nor the differences with migrant entrepreneurs conducting a domestic (namely, non-transnational) business. The paper focuses on the use of social networks by transnational and domestic entrepreneurs, by illustrating the case of Moroccan migrant entrepreneurs in Amsterdam and Milan. The research reveals that transnational entrepreneurs have more heterogenous, geographically dispersed and territorially articulated support networks, which are both more extensive and less dense than those of domestic entrepreneurs. Transnational entrepreneurs exploit their capacity to have contacts with different people in different places. In contrast, domestic entrepreneurs tend to mobilize resources within their dense, homogeneous, and locally based networks.

KEYWORDS:

• Migrant entrepreneurs
• social networks
• transnational entrepreneurship
• transnationalism

Introduction

Networks of social relations generate value for (migrant) entrepreneurs. Social contacts, and the resources available through these, are particularly important for migrant entrepreneurs (Salaff et al., 2003; Sommer & Gamper, 2018). Migrants and migrant entrepreneurs can benefit from social networks as they can acquire tangible and non-tangible resources from their contacts (Portes, 1998; Portes & Sensenbrenner, 1993). For example, they can receive information about business opportunities and ideas for future developments.

This article addresses the kind of social contacts that transnational migrant entrepreneurs employ to access resources and, more in general, receive support to run their business. Transnational migrant entrepreneurs are migrant entrepreneurs who establish a transnational business, namely a business that spans across national borders (Drori et al., 2009). To identify their peculiarities, this article compares transnational entrepreneurs with migrant entrepreneurs engaged in non-transnational business activities, referred to as ‘domestic entrepreneurs’ in this article (see McDougall, 1989). Domestic migrant entrepreneurs are migrant entrepreneurs with a business that focuses exclusively on the domestic market of their country of immigration (Portes et al., 2002 – see also methodology).

Previous studies on social network analysis suggests that the role of social networks depends on the characteristics of the networks, i.e. the nature of the contacts that a person has. One theoretical perspective focuses on dense, rather homogenous, close-knit networks that entail support (Uzzi, 1997). A second perspective views sparse, rather heterogenous, disconnected networks as beneficial for receiving support (Burt, 2001; Granovetter, 1973).

Furthermore, for different business activities (e.g. transnational vs. domestic), different types of social contacts may be used and needed (Stam et al., 2014). Transnational and domestic entrepreneurs might differ in the kinds of social networks they use as well as in the contacts that they receive support and resources from. For example, transnational entrepreneurship is often regarded as a more innovative, profitable and successful form of migrant entrepreneurship, as well as a possible breakout strategy, which entails that different kinds of contacts may be needed (Brzozowski & Cucculelli, 2020; Kariv et al., 2009; Portes et al., 2002; Wang & Liu, 2015).

The literature on both migrant transnationalism and transnational entrepreneurship stresses that migrants strongly rely on their social contacts (Chen & Tan, 2009; Lubbers et al., 2010; Vacca et al., 2018). Previous studies have underlined that networks are the ‘driving force’ of cross-border businesses (Bagwell, 2018; Chen & Tan, 2009; Patel & Conklin, 2009; Portes et al., 2002), but they do not fully clarify the differences between transnational and domestic (non-transnational) migrant entrepreneurs – the few exceptions to this trend are Brzozowski & Cucculelli, 2020; and Portes et al., 2002. In addition, only a few articles employed social network analysis and provided an in-depth analyse the characteristics (composition and structure) of transnational entrepreneurs’ social networks (among others, an exception is Patel & Conklin, 2009).

To contribute to the existing literature, this paper analyses what kind of social contacts transnational migrant entrepreneurs employ to obtain support in order to run their business, and how they differ in this from domestic migrant entrepreneurs. This overall research question refers to two main points of investigation: (1) who provides support to the two types of migrant entrepreneurs (type of social contacts providing support); and (2) who is more likely to provide support to the two categories of migrant entrepreneurs. This article does not focus on the kind of support received by migrant entrepreneurs; rather, it investigates the characteristics of the contacts that provide them with support.

This paper compares two different groups of migrant entrepreneurs from the same country of origin, but residing in two different contexts. It illustrates the results of a research study on Moroccan migrant entrepreneurs with either a transnational or a non-transnational (i.e. domestic) business in Amsterdam or Milan (N = 70). Amsterdam and Milan are two particularly interesting contexts to study and compare as these two cities partially differ in their migration history and migrant population, and this may influence the activities and networks of migrant entrepreneurs in the two cities. The Moroccan group is one of the most significant migrant groups in Amsterdam and in Milan, both in terms of history of migration and size of the population.

This research employs a mixed methods approach by means of face-to-face interviews that combine qualitative questions and quantitative techniques from social network analysis (Coviello, 2005; Nooraie et al., 2020). Following the literature on mixed methods, the research on which this article is based combined elements of qualitative and quantitative research approaches in a single study for the purpose of a deep understanding of the analysed phenomenon (Creswell & Plano Clark, 2011; Johnson et al., 2007).

This paper provides a contribution to the field of transnational migrant entrepreneurs, regarding one example of people that conduct global activities (Portes et al., 2002). Globalization in the form of increased flows of information, communication, capital, goods, and people has created new opportunities for people with relatively limited resources, the so-called ‘globalisers from below’ (Mathews, Lins Ribeiro & Alba Vega, 2012; Smith & Guarnizo, 1998). As stressed by Lins Ribeiro (2012) and Mathews and Alba Vega (2012), globalisers from below are actors with (1) limited power in the global arena, and (2) limited financial capital and resources.

Migrant entrepreneurs are among the actors that the literature considers globalisers from below (Smith & Guarnizo, 1998; Portes, Luis, & Patricia, 1999). Thanks to the opportunities created by globalization (i.e. the advancement of communication technologies and the widespread availability of relatively cheap travel on a large scale), migrants can develop entrepreneurial activities carried out in their country of residence, but with multifaceted links with people abroad (Andreotti & Solano, 2019; Chen & Tan, 2009; Portes et al., 2002; Solano, 2020). In their seminal article, Portes and colleagues (1999) stress the differences between globalisers from above (i.e. corporations and multinational companies) and migrant entrepreneurs as a form of globalization from below: ‘Both make extensive use of new technologies and both depend on price and information differences across [national] borders, but while corporations rely primarily on their financial muscle to make such ventures feasible, immigrant entrepreneurs depend entirely on their social capital. The social networks that underlie the viability of such popular initiatives are constructed through a protracted and frequently difficult process of immigration and adaptation to a foreign society that gives them their distinct characteristics’ (Portes et al., 1999, p. 9). Therefore, migrant entrepreneurs take advantage of the possibilities provided by globalization processes and build on transnational connections to create a competitive advantage. For this reason, it is relevant to focus on the kind of contacts migrant entrepreneurs make use of and receive support from to develop and conduct their transnational business activities.

To study how transnational migrant entrepreneurs, as globalisers from below, employ their contacts for the business, the article compares them with domestic entrepreneurs, as a case of migrants that do not fully seize the opportunities of globalization for their business activity. In what follows, I first introduce the paper’s theoretical approach (social network theory) and previous studies on the topic of (transnational) migrant entrepreneurship and social networks. Then, after presenting the contexts of my research (Amsterdam and Milan) and the research methodology, I illustrate the main findings.

Theory

Personal social networks

The concept of personal social networks refers to the existing social contacts that a person has, as well as the relations with and between these (Crossley et al., 2015). Social contacts may provide the person with access to resources. People, in this case migrant entrepreneurs, can benefit from their social networks as they can acquire tangible and non-tangible resources from their contacts, such as information exchange, provision of free/cheap labour, and financial support (Drori et al., 2009; Portes, 1998; Portes & Sensenbrenner, 1993).

However, the literature stresses that this depends on the characteristics of the networks. Different kinds of networks can have different effects on the support received for the entrepreneurial activity (e.g. whether or not a contact can provide support – Anthias & Cederberg, 2009; Kanas et al., 2011; Patel & Conklin, 2009; Portes & Sensenbrenner, 1993).

In this regard, there are two theoretical perspectives that hypothesize different mechanisms behind the access to support from contacts. The first perspective states that embeddedness in dense, close-knit networks is beneficial for receiving support. These networks are often largely composed by strong ties, namely contacts with whom the person has emotional closeness, e.g. relatives or close friends (Granovetter, 1973; Marsden & Campbell, 1984). In this kind of network, people take advantage of their support from homogenous contacts. In other words, people seem to seek and receive resources from homogenous contacts. Mutual support is favoured by a high level of (emotional) closeness between the contacts and the entrepreneur (Coleman, 1988). Contacts are therefore willing to share information, advice, or financial resources. Similarly, the entrepreneur might be keener to accept when the information or advice comes from those contacts.

In contrast, the second perspective views close-knit connections as a constraint. This view stresses that combining different kinds of contacts may be beneficial to acquire the needed resources. Burt (2001) argues that the more disconnected contacts a person has, the more support the person can access. Contacts in sparse networks are more likely to link different kinds of contacts and different groups of people. Burt (1992, 2001) points to the benefit of having many structural holes in the network. Structural holes refer to people who act as brokers, creating connections between otherwise disconnected contacts. Previous scholars have also pointed to the key support received from weak ties, i.e. contacts that are not very emotionally close (e.g. acquaintances and work-related contacts) (Granovetter, 1973). This second perspective suggests that individuals are more likely to receive support and benefit from having heterogenous contacts in their networks. In contrast, entrepreneurs with highly connected contacts lack access to other groups, and hence – for example – are more likely to receive redundant information and advice compared to others who have access to several different groups. Therefore, different kinds of networks may or may not favour support from contacts. For different business activities and goals (e.g. transnational vs. domestic), different types of social contacts and network characteristics may be needed (Arrighetti et al., 2014; Solano, 2015).

In the next section, I relate the topic of social networks and support received through social contacts to the debate on (transnational) migrant entrepreneurship.

Social networks and (transnational) migrant entrepreneurship

A large body of literature has analysed the determinants and resources that migrants employ to start and conduct a business and the determinants of their entrepreneurial ‘success’ (Rath et al., 2020).

A particular focus has been devoted to migrant entrepreneurs’ social embeddedness and the resources they mobilize in the form of social contacts (Lassalle et al., 2020; Pécoud 2010; Rath et al., 2020; Sinkovics and Reuber, 2021). The literature points to the ability of the entrepreneurs to mobilize resources from networks, within the so-called ethnic community and beyond (Aldrich & Waldinger, 1990; Kloosterman, 2010; Lassalle et al., 2020; Portes & Sensenbrenner, 1993). The need to adapt to the country of immigration creates a certain degree of solidarity among people from the same migrant group (Portes & Sensenbrenner, 1993). Empirical studies on migrant entrepreneurship show that migrants often rely on resources from dense co-national networks in the country of immigration (Anthias & Cederberg, 2009; Cederberg & Villares-Varela, 2019; Kloosterman et al., 2016; Light et al., 1994; Solano, 2015). The co-national group (and the extended family) allows migrant entrepreneurs to mobilize resources (e.g. cheap labour forces, and access to credit) and to access information.

However, support from the ethnic community entails a series of obligations, exchanges and reciprocal favours, as well as limiting access to other relevant information (Bradley, 2004; Gomez et al., 2015; Portes, 1998; Portes & Sensenbrenner, 1993; Shinnar et al., 2011). Over-reliance on migrant networks might constrain the activities of migrant entrepreneurs’ and hamper their business development and sustainability (Arrighetti et al., 2014; Bates, 1994; Lassalle & Scott, 2018; Marger, 2001). To ‘break out’, migrant entrepreneurs require access to a broader range of resources that often go beyond the ethnic community, and may put in place strategies to diversify their networks (Arrighetti et al., 2014; Kanas, van Tubergen, & van der Lippe, 2009; Lassalle et al., 2020).

This seems to apply to transnational entrepreneurs. Transnational entrepreneurship is often regarded in the literature as a more innovative, profitable, and successful form of migrant entrepreneurship, as well as a possible breakout strategy (Brzozowski & Cucculelli, 2020; Kariv et al., 2009; Portes et al., 2002; Wang & Liu, 2015). The results from Santamaria-Alvarez et al. (2018) on Colombian transnational entrepreneurs in the United States suggest a substitution effect between transnational ties and ties in the country of immigration. Here, the authors also show that either transnational entrepreneurs make strong use of their relatives for the business, or they build strategic networks with no strong ties at all.

Nevertheless, there is an extensive body of research that emphasizes the role of (extended) family as well as that of co-national contacts for transnational entrepreneurs (Bagwell, 2015, 2018; Brzozowski et al., 2017; Brzozowski & Cucculelli, 2020; Kariv et al., 2009; Moghaddam et al., 2018; Mustafa & Chen, 2010; Rusinovic, 2008; Solano, 2020; Sommer, 2020; Sommer & Gamper, 2018). For example, Brzozowski and Cucculelli (2020) found that transnational entrepreneurs in Italy also have co-ethnic business ties, and Bagwell (2018) showed that Vietnamese entrepreneurs in London used several contacts and links abroad both from their extended family and in the Vietnamese diaspora to further develop their businesses.

Furthermore, transnational entrepreneurs seem to rely on ‘glocalised’ networks, i.e. networks with both global and local connections (Chen & Tan, 2009). They maintain business activities and contacts with people outside their country of immigration, alongside a certain degree of (local) embeddedness in this country. For example, Patel and Conklin (2009) show that balancing the types and the geographical distribution of contacts provides transnational entrepreneurs with different kinds of resources and opportunities for the business. Sommer and Gamper (2018) found that migrant entrepreneurs with intensive transnational business activities had rather heterogenous and geographically dispersed networks, along with pre-existing strong and weak ties in their country of origin.

Despite this emerging body of literature, deep knowledge on the support networks of transnational entrepreneurs and their peculiarities is still limited. This is mainly due to the scarcity of studies employing social network analysis and comparing transnational entrepreneurs with non-transnational migrant entrepreneurs. Therefore, how these two categories differ is still unclear. The main difference between transnational and domestic entrepreneurs appears to concern the geographical scope of their networks, which is inherently linked to the definition of the two different kinds of migrant entrepreneurs. While support to domestic entrepreneurs seem to come mainly from contacts in their country of immigration, transnational entrepreneurs generally rely on social contacts located both in the country of origin and in the country of immigration, or in other countries (Bagwell, 2015, 2018; Patel & Conklin, 2009; Sommer & Gamper, 2018). Portes et al. (2002), which is the only study comparing the two kinds of migrant entrepreneurs while including an analysis of their networks, observed that a higher number of contacts outside the city of residence (in the country of immigration) decreased the possibility of being involved in domestic entrepreneurial activities. They also found that additional social contacts (network size) increased the probability of running a transnational business.

Background to the study: Moroccan entrepreneurs in Amsterdam and Milan

This research focuses on one national group across two cities in two national contexts. The Moroccan group is one of the most significant migrant groups in Amsterdam and Milan, both in terms of history of migration and size of the population.

Amsterdam and Milan are two particularly interesting contexts to study and compare, because they both play a key economic role in their country (Bontje & Sleutjes, 2007: Mingione et al., 2007).

However, these two cities differ in their migration history and migrant population. This may influence the activities and networks of migrant entrepreneurs in the two cities. Amsterdam and the Netherlands have a longer, 60-year-old history as a migrant destination. In the Netherlands, immigration flows from abroad started after the Second World War, first with arrivals from former colonies (Antilles, Indonesia, and Suriname), and then, from the 1970s onwards, from Mediterranean countries such as Italy, Morocco, and Spain (Lucassen & Penninx, 1997; Bijwaard, 2010). The migrant population of the Netherlands has constantly increased due to further arrivals (mainly) from Morocco and Turkey (Rath, 2009). By contrast, Italy and Milan were characterized by emigration flows for a long time. In the 1970s this trend started to reverse (Colombo & Sciortino, 2004), and since the early 1980s, arrivals from abroad have continued to increase, although only in the 1990s did Italy start face large-scale immigration flows, particularly from nearby Albania and North-African countries. The first Moroccan migrants did not arrive in Milan and Italy until the late 1980s (Colombo & Sciortino, 2004).

As a result, the migrant populations in the two cities are in part different. Both cities have a substantial number of migrant residents, but the number of people of Moroccan background is significantly higher in Amsterdam than in Milan.¹ About 19,000 Moroccan migrants lived in the Milan area in 2019 (0.7% of the total population; source: Istat, 2019), compared to 77,000 Moroccan residents in the Amsterdam area (9% of the total population; source: Centraal Bureau voor de Statistiek, 2019). The relative numbers of migrant entrepreneurs slightly differ in the two cities, with migrant entrepreneurs being 33% of the whole entrepreneurial population in Amsterdam (Rath, 2011), and 29% in Milan (source: Chamber of Commerce of Milan, 2019). There were about 3,100 Moroccan entrepreneurs in the Milan area in 2018 (source: Chamber of Commerce of Milan, 2019), representing 2.5% of all entrepreneurs in the city. While updated data on the national origin of migrant entrepreneurs are not available for Amsterdam, there were about 8,400 Moroccan entrepreneurs in the Netherlands in 2009 (0.6% of all entrepreneurs; source: Chamber of Commerce of Amsterdam, 2009).

Methodology

This research addresses first-generation Moroccan entrepreneurs who have migrated to Milan or Amsterdam. Seventy face-to-face interviews were conducted with participants including transnational (n = 35) and domestic (n = 35) first-generation Moroccan migrant entrepreneurs living in Amsterdam (n = 30) and Milan (n = 40).

Transnational entrepreneurs are migrant entrepreneurs who own a transnational business, namely, ‘individuals who migrate from one country to another, concurrently maintaining business related linkages with their former country of origin, and currently adopted countries and communities. By traveling both physically and virtually, … (they) simultaneously engage in two or more socially embedded environments, allowing them to maintain critical global relations that enhance their ability to creatively, dynamically, and logistically maximize their resource base’ (Drori et al., 2009, p. 1001). Therefore, a transnational business is a business which spans across national borders and entails economic activities linked with other countries, namely outside the country of immigration – e.g. import/export businesses (Chen & Tan, 2009; Drori et al., 2009; Portes et al., 2002). In this paper, I employ the concept of transnational business as opposed to that of domestic business, which in turn is a business that focuses exclusively on the domestic market of this country (Portes et al., 2002). ‘Domestic entrepreneurs’, in this paper, are therefore, migrant entrepreneurs who own a domestic business. It is important to stress that the word transnational, as opposed to domestic, refers to the business and not to the migrants. It is the business that is either transnational or not, rather than the migrant entrepreneur. Indeed, a certain degree of transnationalism characterizes all migrants’ lives, as the literature shows (see for example Vacca et al., 2018).

To distinguish between transnational and domestic (non-transnational) entrepreneurs, I combined the approaches of Rusinovic (2008) and Portes et al. (2002). More particularly, following Rusinovic’s remarks, I opted for a general approach without stressing the fact that transnational entrepreneurs must travel abroad at least twice a year, as in Portes et al., 2002. However, I used a key question asked by Portes et al. (2002), i.e. ‘Is there a relevant part of your business that is related with your country of origin or with other countries outside Italy/the Netherlands?’. The distinction between transnational and domestic entrepreneurs was based on this very question. However, in order to avoid bias linked to self-reporting, the answer to this key question was discussed in-depth during the interviews.

Data collection and sampling strategy

This study adopts a micro-level perspective in that it considers the individual experiences, strategies, networks, and narratives of Moroccan entrepreneurs who have migrated to Milan or Amsterdam. A purposive sample was chosen based on qualitative typologies (Silverman, 2013), which means that entrepreneurs were selected based on the different types of business within the category of Moroccan entrepreneurs with a transnational or a domestic business.

To obtain a comprehensive picture of Moroccan entrepreneurial activities in each city, four sources of participant contacts were used: (1) a list of entrepreneurs provided by the Milan Chamber of Commerce, with the indication of the business sector (e.g. import/export) and a short business description (no list was available for Amsterdam); (2) contacts from Moroccan associations with a significant role in the Moroccan communities of Amsterdam and Milan (e.g. Moroccan businesses and Islamic cultural associations); (3) entrepreneurs’ business cards and advertisement materials left in shops and stores; (4) Moroccan shops and stores that were visible in ethnic and central neighbourhoods of Amsterdam and Milan. Sometimes, when walking around city areas with a high density of Moroccan businesses, I noticed a particularly interesting business for the research purpose (e.g. import/export) and I simply asked the owner for an interview.

The research employs a mixed-method approach by means of face-to-face interviews that combine qualitative questions and quantitative techniques from social network analysis. Following the literature on mixed methods, the research combined qualitative and quantitative elements in a single study, at all stages – from the fieldwork to the data/information analysis and the illustration of findings –, to answer the same research question (Creswell & Plano Clark, 2011; Johnson et al., 2007). This is particularly fruitful when applying a social network approach, as it allows for a deep understanding of the mechanisms behind the network characteristics identified through quantitative techniques (Nooraie et al., 2020).

From September 2013 to November 2014, I conducted 70 face-to-face interviews that combined qualitative questions and quantitative questions regarding the interviewees’ networks. As already done in existing studies applying social network analysis and mixed methods (Coviello, 2005; Nooraie et al., 2020), the interviews were based on a structured interview guide that alternates open-ended questions with social network analysis items.² All the questions were asked orally in discursive fashion, to encourage fluidity and collect more detailed information. The interviews lasted from one hour and a half to three hours, and they were conducted on the business premises. The interviews were conducted in either the official language of the destination country (Italian or Dutch) or in English, when the respondents opted for it. Besides socio-demographic information, questions regarding entrepreneurial experience, daily working practices, business contacts, and resources used were investigated.

Qualitative questions (e.g. ‘Please describe in detail the activity of your business, and in particular, activities connected with, or located in, your country of origin or in other countries’) were useful for a detailed understanding of the participants’ entrepreneurial activities and the help they received.

To collect data regarding the entrepreneurs’ networks, I made use of social network analysis for ego-networks (Crossley et al., 2015), namely ‘Personal Network Analysis’ (PNA). PNA is the part of social network analysis focusing of personal networks, most importantly those where all the persons mentioned share the characteristic of being in direct connection with the interviewee. PNA was used here in order to understand the entrepreneurs’ contacts, their characteristics, as well as the help received from these categories of contacts.

To collect network data, interviewees were asked the following question, to elicit mention of their contacts,

Would you please give us the names of 30 people (alters) that you know and who know you, with whom you have had any contact in the last two years, and that you could still contact if you needed to?

Following the relevant literature, I asked for a fixed number of persons (30). Personal networks of less than 20 alters are more likely to produce a biased picture of the network, while a number between 30 and 60 alters should be asked for in order to obtain accurate information of a person’s network and its characteristics such as composition and density (McCarty et al., 1997). In this regard, I made a distinction between business and non-business-related networks (Mollenhorst, Volker, & Flap, 2008; Schutjens & Völker, 2010): ‘business network’ refers to persons involved in the business, and ‘non-business-related network’ refers to the persons closest (in terms of personal relationships) to the interviewees. I made this difference in order to understand the interplay between these two types of contacts for the business. Interviewees were asked to try to balance the names they gave between non-business-related (e.g. relatives) and business-related contacts (15 non-business contacts and 15 business contacts). The choice of asking the entrepreneur to try to balance the two categories was made to avoid interviewees just listing all their relatives without thinking about other relevant contacts.

To avoid any bias in the definition of the business network and to understand whether the people in the network were involved in the business (business network), I asked, ‘Do you have contacts with him/her for business purposes?’. The following question served to find out whether a contact provided support for the business (support network): ‘Does he/she provide support for the business?’. Furthermore, a question regarding the kind of support provided was also asked (‘What kind of help does he/she mainly provide?’).³ Based on some literature on the topic (Rath et al., 2020), ‘I distinguish between three different kinds of support’: 1. Financial support (e.g. a loan); 2. Working help; 3. Providing information. I also included the category ‘other’, but few interviewees opted for this, and their answers fell under the three pre-determined answer options. The kind of support provided was also discussed in detail during the entire interview.

People for whom the interviewee answered ‘yes’ to the previous two questions (‘business contacts’ and ‘providing support’ questions) were included in the business network, while the support network was composed of people providing support (based on the ‘provide support’ question). The content of these answers was discussed during the interviews to clarify for the respondents the meaning of the questions asked. Therefore, while the entire network has a fixed number of contacts, the number of contacts in both the business and support network varies.

For each contact named by the interviewee, I collected the following information: gender; place of birth; place of residence; relationship with ego; contacts for the business; support provided for the business (see also the ‘Analyses’ section). Furthermore, to assess network structure (density), I also investigated whether the contacts knew each other (‘Do these two persons know each other?’). The interviewees were asked to answer to this question for each pair of persons. This is a normal and widely used procedure in personal network analysis (McCarty, Lubbers, Vacca, & Molina, 2019).

Sample

Most research participants were male (n = 54), middle aged (about 40 years old), with a medium-high level of education (see Table 1). This is in line with the sex, age and education distribution of the most recent data from OECD (OECD, 2010) about migrant entrepreneurs in Italy and the Netherlands at the time of the research study. I interviewed entrepreneurs with both ethnic (n = 27) and non-ethnic/mainstream (n = 43) businesses. Following the literature on the topic (Ambrosini, 2012), ‘ethnic market’ refers to a market in which the products are mainly for a clientele of co-nationals and other immigrants. ‘Mainstream market’ refers to a market addressing all types of clients (natives, co-nationals and other immigrants). Participants usually owned small businesses, with one or two employees (64%). Transnational and domestic entrepreneurs have rather similar profiles. The only main difference is that the sample of domestic entrepreneurs is made up of more businesses in the mainstream market.

Table 1. Sample characteristics.

Support	Whole sample	Transnational e.	Domestic e.
Amsterdam	30	15	15
Milan	40	20	20
Men	54	26	28
Age (median)	43	39	44
Higher Education (high school or more)	51	27	24
Mainstream business	43	17	26
Ethnic business	27	18	9
Number of employees (median)	2	1	2
Year of start-up (median)	2008	2009	2007
Total (N)	70	35	35

Various activities are included as domestic businesses: retailers (e.g. minimarkets), cafes, bakeries, greengroceries, hairdressers, etc. As for transnational businesses, they include the following main types of business: retail or wholesale of imported products; import and/or export businesses; consultancy agencies (e.g. mediation and counseling for individuals and companies).

Analyses

The collection of network data led to the creation of an entrepreneur(ego)-level dataset (N = 70) and a contact(alter)-level dataset (N = 1,988). Out of the collected alters, 963 contacts composed the business networks (49% of the total number of alters), while 1025 (51%) were non-business-related contacts (e.g. relatives not involved in the business at all. Out of the business-related contacts, 490 contacts provided support to the entrepreneurs, thus being part of the support networks (25% of the total number of alters; 51%of the business-related alters)). The other 473 business-related contacts did not provide any kind of support to the entrepreneurs.

The entrepreneur-level dataset contains interviewees’ information, including network-related information (e.g. network size, share of Moroccan contacts). The contact-level dataset contains information on the alters, alongside data on the related ego.

These two datasets were used to perform bivariate and multivariate analyses. In the analyses, I looked at both network composition and structure. Network composition refers to the kind of contacts that entrepreneurs have, and their relationship with these contacts, while network structure concerns the number of contacts that a person has and the links between contacts. Therefore, I looked at the following independent variables:

• Network size. Size refers to the number of contacts included in a network.

• Network density. Density represents the proportion of actual relationships in the network compared to the maximum number of potential relationships. In other words, if every contact mentioned had a relationship with every other contact mentioned, then density would be ‘1’; if no one was connected to any other person in the ego-network, density would equal ‘0’.

• Relationship between the entrepreneur and the contact. I asked about the relationships and distinguished between the following categories: close relative/partner, other relative; friend; acquaintance; work-related contact. Parents, children, siblings, and grandparents were considered close relatives.

• Tie strength. Based on previous literature (Marsden & Campbell, 1984; Mustafa & Chen, 2010), I included the following contacts as strong-tie categories: (1) close relatives; (2) friends; (3) people interviewees qualified as either close or very close. This was done to obtain a measure of strength based on both kinship and emotional closeness, in line with other studies on migrants’ social networks (Lubbers et al., 2010; Molina et al., 2015; Vacca et al., 2018), to establish point 3 (people interviewees qualified as either close or very close). I have asked the following question for each contact: ‘How close do you feel to this person? (1- Not at all, I don’t feel close at all to this person; 5- I feel very close- to this person)’. The closeness here refers to kinship and not to friendship. During the interviews, in order to avoid mistakes, interviewees were asked to specify the kinship (e.g. parent, cousin) and then to decide if the relative was ‘close’ or ‘other’. Following Mustafa and Chen (2010), relatives from extended families (i.e. other relatives) were not automatically considered strong ties. This is also supported by the data. In the sample, only 45.3% of other relatives were considered close or very close (in terms of friendship) by the interviewees. In contrast, 88.9% of close relatives (in term of kinship) were considered close or very close (in terms of friendship). The data support the inclusion of ‘friends’ as strong ties, since 93% of friends were considered close or very close (in terms of friendship) by the interviewees.

• Groups. Entrepreneurs’ contacts can also be distinguished in terms of their country of origin and their country of residence. By combining these two pieces of information, the following groups of contacts can be identified (see Lubbers et al., 2010; Molina et al., 2015; Vacca et al., 2018): natives of the country of immigration; Moroccan migrants in the country of immigration; other migrants living in the country of immigration; Moroccans living in Morocco (Moroccan origins); Moroccan diaspora persons; international persons, namely people of any nationality (other than Moroccan) living in countries that are neither Morocco nor the interviewee’s country of immigration.

• Place of residence (country and city) of the contact. Information on where the contacts lived (city and country or residence) was collected.

• When the entrepreneur and the contact had met (before or after the business start-up). I asked about when the entrepreneurs met their contacts, distinguishing between before and after the business start-up.

As for the analyses conducted, I first employed nonparametric bivariate tests (the Wilcoxon rank-sum test and chi-square statistic) to compare the support networks’ characteristics of transnational and domestic entrepreneurs. I also ran multilevel logistic regressions to check what kind of contacts were more likely to support transnational and domestic entrepreneurs. I employed a multilevel model, since the data had a nested structure consisting of multiple ties per responding entrepreneur. The dependent variable was whether the contact provided support for the business (dummy variable: ‘Does he/she provide support for the business?’).

The quantitative analysis is integrated with an examination of the qualitative parts of the interviews, which are used to investigate the reasons behind the differences between transnational and domestic entrepreneurs observed in the data, as well as their contacts’ roles.

Findings

In what follows I focus on the support networks, as part of the interviewees’ business networks (see methodology). I first analyse the kind of contacts interviewees have in their support networks (i.e. contacts providing support), by comparing the characteristics of transnational and domestic supporting networks’ (who provide support). Second, I focus on what kinds of social contacts are more likely to provide support to both types of migrant entrepreneurs, among the contacts that are somehow related to the business activity.

Entrepreneurs’ support networks

In this section I focus on the kind of contacts providing support. This section summarizes the main characteristics of transnational and domestic entrepreneurs’ support networks, namely the entrepreneurs’ contacts, and whether these differences are statistically significant, see Table 2 and 3. Many differences emerge in the characteristics of the support networks of transnational and domestic entrepreneurs.

Table 2. Support network characteristics: composition and structure.

Network characteristics	Median	Mean	Z	p
Network size
Domestic entrepreneurs	8	6	−2.244	0.03
Transnational entrepreneurs	9	1.34
Network Density
Domestic entrepreneurs	0.5	0.5	3.657	0.00
Transnational entrepreneurs	0.2	0.3
Share (%) of work-related contacts
Domestic entrepreneurs	46	43	−1.289	0.20
Transnational entrepreneurs	60	53
Share (%) of relatives
Domestic entrepreneurs	20	33	−2.427	0.92
Transnational entrepreneurs	22	27
Share (%) of strong ties
Domestic entrepreneurs	41	51	−1.109	0.27
Transnational entrepreneurs	67	59
Share (%) of contacts met before the business start-up
Domestic entrepreneurs	59	60	−0.380	0.70
Transnational entrepreneurs	60	63
Share (%) of contacts living abroad
Domestic entrepreneurs	0	3	−5.726	0.00
Transnational entrepreneurs	39	42
Share (%) of contacts living outside the interviewee’s city of residence
Domestic entrepreneurs	86	82	4.889	0.00
Transnational entrepreneurs	33	38

Note. The table displays the median value of network characteristics among transnational and domestic entrepreneurs. The table displays the result of the Wilcoxon rank-sum tests, which is a nonparametric bivariate test employed here to analyse whether the characteristics of the support networks of transnational and domestic entrepreneurs were statistically significant. Z and p are the test statistic and p-value, respectively, from Wilcoxon rank-sum tests. Bold p-value indicates test significance at 5% level.

Table 3.  

Support network characteristics: share of alters in the different groups of contacts (based on country of origin and country of residence).
Group	Support network
	Transnational e.	Domestic e.
Natives	32.4	43.3
Moroccan migrants	20.9	42.4
Moroccan Origins	16.4	1
Other immigrants	4.9	11.8
Diasporas	5.2	0
Internationals	20.2	1.5
Total (%)	100	100
Total (N)	287	203
Chi-square test(df)	X^2(5) = 101.911, p = 0.000

Note. The table displays the distribution for each group of contacts of transnational and domestic entrepreneurs. The Chi-square test is used to determine whether there is a statistically significant difference between expected and observed frequencies in one or more categories (in this case, transnational or domestic entrepreneurs) of a contingency table.

First, the network structure of transnational and domestic entrepreneurs differs in density and size. Transnational entrepreneurs have slightly larger and less dense business support networks than do domestic entrepreneurs. Transnational entrepreneurs have on average eight people supporting them, while domestic entrepreneurs have five supporting contacts in their support network. This difference is statistically significant, as a Wilcoxon rank-sum test shows (Table 2).⁴ Transnational entrepreneurs’ networks (Mdn = 0.2) are less dense than those of transnational entrepreneurs (Mdn = 0.5). This difference is statistically significant (see Wilcoxon rank-sum test in Table 2 and 3).

Second, transnational entrepreneurs have more diversified networks than domestic entrepreneurs in terms of background (national origin) and place of residence (Table 2 and 3), as they receive support from different groups: Natives, Moroccan migrants, Moroccans living in Morocco (Moroccan origins) and Internationals (non-Moroccan people living in a third country that is neither the country of immigration nor Morocco). In contrast, the supporting contacts of domestic entrepreneurs’ are almost exclusively made up of natives and Moroccan migrants in their country of immigration.

Third, for domestic entrepreneurs, the data suggest a major concentration of supporting contacts in the country of immigration (Table 2 and 3). Transnational entrepreneurs have a higher share of supporting contacts living abroad (Mdn = 39%) compared to domestic entrepreneurs (Mdn = 0%), as the Wilcoxon rank-sum test confirms (Table 2 and 3). Furthermore, the networks of transnational entrepreneurs (Mdn = 33%) have a smaller share of supporting contacts living in the entrepreneur’s city of residence than those of domestic entrepreneurs (Mdn = 86%), see (Table 2 and 3). The greater geographic dispersion of transnational entrepreneurs’ supporting contacts is somehow linked to the definition of the two types of entrepreneurs (see Methodology). Entrepreneurs were classified as transnational if they had a relevant part of their business outside their country of immigration, whereas domestic entrepreneurs were migrant entrepreneurs who focused only on their country of immigration’s domestic market (see methodology). Therefore, this difference was expected. However, having a relevant part of the business abroad does not automatically entail having a high(er) number of contacts providing support from abroad. In fact, even a single link (e.g. a supplier) might lead to having a key part of the business outside the country of immigration. Furthermore, the literature has underlined that all migrants are transnational and have relatives, friends and other contacts scattered all around the world (Solano et al., 2020; Vacca et al., 2018). Therefore, even for domestic entrepreneurs, support could, in theory, have come from abroad.

In contrast, not only do transnational entrepreneurs have fundamental contacts abroad (in terms of importance, as asked during the interviews), these contacts are also relevant in numerical terms (i.e. the number of business contacts abroad). Furthermore, domestic entrepreneurs do not receive support from people located abroad, or even outside the city where they live.

Fourth, both transnational and domestic entrepreneurs strongly rely on the support of contacts that they met before they started the business. Transnational entrepreneurs also have almost two-thirds of their supporting network made up of pre-business contacts (60%). The same applies to domestic entrepreneurs (59%).

Lastly, the business support networks of both transnational and domestic entrepreneurs have many exclusively work-related contacts, i.e. people who have contacts with the entrepreneur exclusively for the business. While transnational entrepreneurs have a majority of work-related contacts (Mdn = 60%), such work-related contacts make up just under half of the supporting contacts for domestic entrepreneurs (Mdn = 46%). However, this difference is not statistically significant (see Wilcoxon rank-sum test in Table 2 and 3). Despite this, relatives are important for both categories of entrepreneurs, as the networks of transnational and domestic entrepreneurs have a similar number of relatives, 22% and 20% respectively. Family relations remain important and very much ‘present’ even when the relatives live far away. However, the relevance of strong ties, meaning people the entrepreneur are emotionally close to, seem higher for transnational entrepreneurs (Mdn = 67%) than for domestic entrepreneurs (Mdn = 41%), although this difference is not statistically significant, according to a Wilcoxon rank-sum test (see Table 2 and 3).

During the interviews, transnational entrepreneurs clarified why relatives and, more generally, strong ties were so important for their business. They tend to rely on strong ties to manage the part of the business that is located furthest from the entrepreneur’s place of residence, or to obtain key information regarding opportunities located outside the context of immigration where they started their business:

My brother lives in Barcelona. He has direct contacts with the suppliers there. When we started, he went directly to the central market in Barcelona, and he spoke with them in order to broker some agreements. He usually buys the fruit; he arranges everything and then he sends the fruit to Milan. (M06)

I send part of the clothes I buy here to Morocco. Then, my brother sells them to shops over there. (M01)

My friend from the Emirates supports me in selling my products in the Gulf. He gives me tips; he introduces me to some clients, etc. (A22)

I have one supplier in Morocco. I chose that one because the company is quite famous, and some friends knew it personally. (M14)

Some of my relatives have a similar business in another part of Morocco. Then one of my father’s neighbours also gave me some information and he helped me to start a business in this field. (M11)

The above quotes also show the role of strong ties as brokers (related to structural holes, see Burt, 2001), i.e. persons who connect different groups and different kinds of contacts. These contacts allowed transnational entrepreneurs to develop their business and reduce the initial investments, both in terms of financial resources – for example, they would likely need to spend significant sums on travel without the help of strong ties – and in terms of acquiring key contacts – as stressed in the quotes.

For example, J. (A20), a transnational entrepreneur living in Amsterdam, imports and sells Moroccan fabrics. A key contact for J. is his brother, who has a similar business and who supported him with information regarding possible suppliers to contact. N. (M14) imports and retails Moroccan foods and herbs. His business network is mainly made up of strictly work-related contacts. The only strong tie in his business network is L., a friend who introduced N. to key suppliers.

All in all, transnational entrepreneurs and domestic entrepreneurs differ in the type of contacts they receive support from (support network – see also online appendix for examples of the social networks of the two types of migrant entrepreneurs). Transnational entrepreneurs display rather sparse, heterogenous, and geographically dispersed supporting networks. This suggests that they bridge different kinds of contacts to obtain support for their business. Nevertheless, transnational entrepreneurs still rely on relatives and strong ties, who often act as brokers allowing the entrepreneurs to connect with other key contacts. In contrast, domestic entrepreneurs have more dense, homogenous, and geographically concentrated supporting networks.

Who is more likely to provide support?

Transnational and domestic entrepreneurs do not differ only in their support networks, but also in where they are more likely to receive support from (among their business-related contacts).

The basic unit of analysis in our study is the relationship between the ego (the entrepreneur) and an alter (the contact). Since there are multiple relationships per entrepreneur, the data are characterized by a nested structure. Most statistical techniques assume independence between observations. This restriction has hampered studies on social networks since they could only focus on the relational level or the network level (Snijders & Bosker, 1999). To deal with the nested structure of the data, I apply a multilevel statistical model. This is a very common and widely used procedure in social network analysis (Snijders & Bosker, 1999; Vacca, 2018).

Table 4 display the results of the multilevel logistic regression analysis regarding the kinds of networks and contacts that are more likely to provide any kind of support to transnational and domestic entrepreneurs. Correlations between variables were checked (results not reported here) and no high correlations emerged.⁵

Table 4. Multilevel logistic regression. DV: whether the contact provides support (1) or does not (0).

Transnational entrepreneurs’ sample	β		SE	OR
Ego level
City (0 – Milan; 1 – Amsterdam)	0.09		0.42	1.09
Business (0 – ethnic; 1- mainstream)	−0.25		0.47	0.78
Sex (0 – Female; 1- Male)	0.17		0.47	1.19
Education (0 – low; 1- high)	0.36		0.52	1.43
Business Network Size	0.11	***	0.04	1.12
Business Network Density	−0.6	p = 0.09	0.35	0.55
Alter level
Sex (0 – Female; 1- Male)	1.08	****	0.23	2.95
Living immigration city	−0.46	***	0.18	0.63
Strong tie	0.59	***	0.22	1.8
Moroccan contact	−1.01	****	0.2	0.36
Contact known before business start-up	0.54	**	0.23	1.72
Constant	−4.81	***	1.55	0.01
N observations	900
N entrepreneurs	30
SD (u)	0.86
Log likelihood	−477.45793
AIC	981.2296
BIC	1043.661
Wald (df)	63.39(11)	***
Domestic entrepreneurs’ sample	Β		SE	OR
Ego level
City (0 – Milan; 1- Amsterdam)	−0.68	*	0.36	0.51
Business (0 – ethnic; 1- mainstream)	−0.92	***	0.38	0.4
Sex (0 – Female; 1- Male)	−0.81		0.46	0.44
Education (0 – low; 1- high)	−0.89	***	0.35	0.41
Business Network Size	0.04		0.03	1.04
Business Network Density	−0.53	**	0.22	0.59
Alter level
Sex (0 – Female; 1- Male)	0.61	***	0.23	1.83
Living immigration city	0.66	***	0.23	1.94
Strong tie	−0.18		0.2	0.83
Moroccan contact	−0.25		0.21	0.78
Contact known before business start-up	−0.41	*	0.21	0.67
Constant	0.98		1.18	2.67
N observations	930
N entrepreneurs	31
SD (u)	0.66
Log likelihood	−411.28831
AIC	848.5766
BIC	911.434
Wald (df)	55.59(11)	***

*p < 0.06; **p < 0.03; ***p < 0.01; ****p < 0.001.

The size of the business network has a positive effect in the sample of transnational entrepreneurs (β = 0.11; p < .01), while the effect is not significant for domestic entrepreneurs. For transnational entrepreneurs, an additional contact in their network increases the chance of receiving support by 1.12 times (see odds ratios, OR column in Table 4).

Density has a significant negative effect for domestic entrepreneurs (β = −0.53; p < .001), while it is weakly significant for transnational entrepreneurs (β = −0.59; p < .1). For domestic entrepreneurs, a higher level of density halves the chances of receiving support (OR = 0.6), similarly to transnational entrepreneurs (OR = 0.55).

Transnational entrepreneurs seem to seek support beyond the ‘ethnic’ community, as they are less likely to be supported by Moroccan contacts (β = −1.01; p < .001). This means that the odds of a Moroccan contact providing support are about half as likely (OR = 0.4) as the odds of a non-co-national contact providing support. The effect is not significant for domestic entrepreneurs.

Opposite effects emerge when it comes to the location of the contacts. Contacts living in the entrepreneur’s city of immigration are less likely to provide support to transnational entrepreneurs (β = −0.46; p < .01) and more likely to provide support to domestic entrepreneurs (β = 0.66; p < .01). This implies that the probability that a contact living in the same city as the entrepreneur will provide support are more than one and half times higher (OR = 1.6) for domestic entrepreneurs, and lower (OR = 0.64) for transnational entrepreneurs, compared to when the contact lives outside the city.

Diverging effects also appear regarding the effect of pre-business contacts, namely contacts met before the start-up of the business. Pre-business contacts are more likely to provide support in the case of transnational entrepreneurs (β = 0.54; p < .03), while they are less likely to do so when the entrepreneur has a domestic business (β = −0.43; p < .06). Therefore, the odds of a pre-business contact providing support are more than one and half times higher (OR = 1.7) for transnational entrepreneurs. By contrast the chance that a pre-business contact will provide support to a domestic entrepreneur are lower: only 0.7 times compared to the probability that a post-business contact will do so. During the interviews, transnational entrepreneurs stressed the key relevance of pre-existing contacts several times. For example, S. (A26), who sells decorative flowers to luxury hotels and restaurants in the Arabian Peninsula, underlined that ‘some [key contacts who support the business] are friends or people I met before my company and my ideas, and they supported me as investors’. Similarly, A. (M40) has a consultancy business helping companies enter the market of MENA countries. He stresses the importance of pre-business contacts: ‘Many people I met before the business support me, both Italians and other immigrants. Most are friends but there are also acquaintances. They simply advise people who need my services to come here’.

Furthermore, transnational entrepreneurs are more likely to receive support from strong ties (β = 0.59; p < .01), i.e. the odds that a strong tie will provide support are almost double (OR = 1.8) those when the contact is a weak tie. The effect is not significant for domestic entrepreneurs. The relevance of strong ties for transnational entrepreneurs is explained during the interviews. Since transnational entrepreneurs need to manage businesses that span across national borders, they need people they can trust to support them for the business. For example, M. (M02) decided to set up a part of the clothes production for her fashion business in Morocco. Her father, who still lives in Morocco, is supporting her by verifying and picking up some of the products from the factory that manufactures the clothes: ‘Sometimes he goes to pick up a package or to check that everything is OK. For this job I needed someone I trusted’. Similarly, J. (A20), who owns a shop selling Moroccan furniture and fabrics, ‘got a lot of information about the business from friends in Morocco’.

All in all, transnational entrepreneurs are more likely to receive support from bigger and less dense networks, contacts outside their city of residence and contacts that are not co-nationals. This does not prevent them from being supported by emotionally close contacts (strong ties) and contacts they met before the business start-up, since the multilevel regression shows that transnational entrepreneurs are more likely to be supported by them.

As for domestic entrepreneurs, although they tend to rely on rather dense, homogenous networks (see previous section), it seems that they do not fully benefit from homogeneity and density. Domestic entrepreneurs are less likely to receive support when contacts are part of a dense network and when they met the contacts before the business start-up.

Discussion and conclusion

This article addresses the social contacts of transnational and domestic migrant entrepreneurs. It offers new insights on the use of social networks by ‘globalisers from below’ (Portes & Sensenbrenner, 1993; Serra del Pozo, 2002). The debate on globalization from below stresses the role of people in the globalization process (Mathews and Vega, 2012), as well as how these actors take advantage of the possibilities provided by globalization processes and build on transnational connections to create a competitive advantage by means of relatively limited resources (Portes & Sensenbrenner, 1993; Smith & Guarnizo, 1998). Their case seems particularly interesting in order for us to understand how ongoing global processes – increased possibilities for international communication and travel – shape migrants’ entrepreneurial choices and how migrant entrepreneurs take advantage of these processes. This article shows how they use their contacts as a resource to conduct their businesses.

This article’s findings show the key role of migrants’ social networks to conduct transnational business activities. Interviewed transnational entrepreneurs would not have been able to run their business without the help of such contacts. Social contacts allow migrant entrepreneurs to conduct transnational business activities, and to be linked to several places at once.

This shows how globalisers from below, with relatively limited (financial) resources, are enabled to carry out activities that span across national borders thanks to key contacts. These contacts allow transnational entrepreneurs to develop their business and reduce their initial investments, both in terms of financial resources and in terms of accessing needed information. These entrepreneurs would not often have had the financial resources to start or to internationalize their business without their contacts. Therefore, the case of transnational entrepreneurs shows how globalisers from below mobilize the resources that they have in order to overcome their ‘subordinate’ position in the global arena. In other words, their power lies in their social networks.

This article offers a contribution to the existing literature on transnational entrepreneurs. This literature suggests that, compared to domestic entrepreneurs, transnational entrepreneurs tend to rely on support from more heterogenous and geographically dispersed contacts, with a higher share of weak ties (Bagwell, 2018; Chen & Tan, 2009; Patel & Conklin, 2009; Solano, 2020; Sommer & Gamper, 2018). However, given that almost no study on transnational entrepreneurs has compared them with non-transnational migrant entrepreneurs, it was not entirely clear what the main differences between them might be (Portes et al., 2002; Rath et al., 2020). By comparing transnational and domestic entrepreneurs, this article clarifies the differences between transnational and non-transnational migrant entrepreneurs when it comes to their social networks and the support they receive from their social contacts.

Compared to domestic entrepreneurs, transnational entrepreneurs have more heterogenous, geographically dispersed, and territorially articulated networks. The business networks of transnational entrepreneurs are also (slightly) larger and less dense than those of domestic entrepreneurs. Interviewed transnational entrepreneurs have more strong ties in their support networks than do domestic entrepreneurs. The heterogenous networks of transnational entrepreneurs do not prevent them from having strong ties in their supporting networks, consistently with previous studies showing that personal contacts contribute to the intensification of transnational business activities (e.g. Brzozowski et al., 2017; Brzozowski & Cucculelli, 2020). The role of these strong ties for transnational entrepreneurs is crucial, as they contribute to managing the foreign side of their business and provide them with key information on opportunities to expand (e.g. abroad). Strong ties can represent a means to link with weak ties operating in a wider social context, e.g. outside the country of immigration (Jack, 2005).

Transnational and domestic entrepreneurs differ not only in their support networks, but also in the types of contacts that they benefit from by having these contacts somehow involved in their business (business network) – namely, the contacts that are more likely to provide support. The regression models regarding what business-related contacts are more likely to provide support display interesting results if compared to the results on the support network. Although domestic entrepreneurs seem to rely on rather dense, homogenous networks, while transnational entrepreneurs rely on more sparse, heterogeneous networks, they do not necessarily benefit from larger networks (transnational entrepreneurs) or denser network (domestic entrepreneurs). The multilevel regression analysis shows that contacts are less likely to support either transnational or domestic entrepreneurs when they are part of denser (business) networks. Contacts are more likely to support transnational entrepreneurs when they are part of larger networks. Therefore, the results suggest that, overall, (business) network size tends to be beneficial for receiving support for migrant entrepreneurs, while (business) network density seems to be detrimental for support.

Furthermore, transnational entrepreneurs receive support from strong ties and pre-business contacts, while pre-business contacts are less likely to provide support to domestic entrepreneurs. This finding might seem to be in contrast with the fact that transnational entrepreneurs bridge different types of contacts to obtain support for the business. Actually, this reveals that transnational entrepreneurs were more able to mobilize already existing resources in their network than were domestic entrepreneurs.

My research confirms, but also partially contrasts, some of the findings from the existing literature. It confirms that, in contrast to transnational entrepreneurs, domestic entrepreneurs do not receive support from people located abroad or, even outside the city where they live. A similar result emerged from the comparative research of Portes et al. (2002), who underlined that domestic entrepreneurship is a form of ‘economic path engaged in by immigrants whose ties do not reach beyond the local community’ (p. 289). Results also show that transnational entrepreneurs combine, and somehow balance, heterogenous contacts, e.g. both local and transnational contacts, as previous studies already suggested (Brzozowski & Cucculelli, 2020; Chen & Tan, 2009; Patel & Conklin, 2009; Sommer & Gamper, 2018).

The literature considers transnational entrepreneurship as a more innovative, more profitable, and more successful form of migrant entrepreneurship (Brzozowski & Cucculelli, 2020; Portes et al., 2002; Wang & Liu, 2015). My findings seem to support this by showing that transnational entrepreneurship represents a breaking-out strategy, and that transnational entrepreneurs follow the ‘breaking-out’ process underlined in the literature (Arrighetti et al., 2014) by connecting with a wide range of contacts.

However, this article’s findings partially contradict Santamaria-Alvarez et al. (2018), who show that either transnational entrepreneurs make strong use of their relatives for their business, or they build strategic networks with no close ties. My findings suggest that they often combine the two strategies and that close ties can help them connect with key business contacts. Therefore, the substitution mechanism illustrated by Santamaria-Alvarez et al. (2018; see also Brzozowski & Cucculelli, 2020) is only confirmed for the geographical distribution, but not for the kinds of relationships, that transnational entrepreneurs have with their contacts who provide support.

Beside its contributions, the article has some limitations, which pave the way for further research. The main limitation is that it addresses one specific group of migrants. As underlined by Portes et al. (2002), different national groups can develop different transnational and entrepreneurial activities. Therefore, other studies should compare various national groups of migrants. Furthermore, this study employs a mixed methods approach applying quantitative analyses to a small number of migrant entrepreneurs. Further studies may either go into more detail regarding the mechanisms underlying the support received by transnational entrepreneurs – by using ethnographic and qualitative methods –, or they could apply the same analytical approach to a larger sample – by means of quantitative analysis.

Nevertheless, this article contributes to the study of the social networks of migrant entrepreneurs by illustrating that different types of social contacts are used by migrants for different business activities. It shows that domestic entrepreneurs tend to mobilize social resources within their dense, homogeneous, and locally based, networks, relying on bonding as a networking strategy to receive support. In contrast, transnational entrepreneurs bridge different people in different places, and have sparse, heterogeneous networks.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by a doctoral scholarship from the Italian Ministry of Education and a mobility scholarship from the European Commission (Erasmus Placement/Erasmus+Traineeship). Furthermore, Fondazione Roberto Franceschi Onlus and Fondazione Isacchi Samaja (Young Professional Grant – 2013 edition) provided the financial support for the fieldwork on which this article is based.

Notes on contributors

Giacomo Solano

Giacomo Solano is Assistant Professor in Migrant Inclusion at the Nijmegen School of Management, Department of Economics and Business Economics. He is affiliated to the Radboud University Network on Migrant Inclusion (RUNOMI), and he cooperates with the Global Data Lab. He holds a PhD in Social Sciences from the University of Amsterdam and University of Milan-Bicocca (joint degree). His research interests include social and labour market integration of migrants, migrant entrepreneurship, comparative integration policies, social dynamics in developing countries and social network analysis. In the past, he worked for the Migration Policy Group (MPG) as Head of Research, the European Commission (DG Employment), the International Organization for Migration (IOM), and the Eindhoven University of Technology (as post-doc researcher).

Notes

1 All Dutch figures include both first- and second-generation migrants. Italian statistics consider only people born outside Italy.

2 The entire interview guide can be consulted here: https://pure.uva.nl/ws/files/4505986/170994_09_Appendix.pdf

3 Given that this article does not focus on the kind of support received, but rather on the contacts that provide support, results linked to this question are not reported here (see Solano, 2016).

4 This difference holds when other business or individual characteristics – business type (ethnic/mainstream),city, sex, and education – are controlled for in a regression model, although the resultwas weakly statistically significant (β = 2.53; p = 0.09). I also ran a set of regression analyses to confirm the other results. This was done each time a Wilcoxon rank-sum test was conducted. The results of the regression analyses, which are not reported here, always confirmed the result of the test.

5 I decided not to include the variable ‘contact living abroad’ in the model for several reasons: having key contacts with people living abroad was part of the definition of transnational entrepreneurs; living in the immigration city and living abroad are very highly correlated (r =  −0.63, p < 0.01). I ran the same model with the variable ‘contact living abroad’ anyway, without the ‘city of immigration’ variable: 1) the overall results were confirmed, except for ‘contact known before the business start-up’ which became very weakly significant (p = 0.09); 2) as expected, transnational entrepreneurs are more likely to receive support from contacts living abroad (β = 0.53; p < .06), while this is less likely for domestic entrepreneurs (β = 0.11; p < .001).