Understanding the gender gap in immigrant entrepreneurship: a multi-country study of immigrants’ embeddedness in economic, social, and institutional contexts

Research shows that women around the world are, in general, less likely to engage in entrepreneurship than men (Kelley et al. 2015; Minniti and Naudé 2010; Elam and Terjesen 2010; Verheul et al. 2006). The general entrepreneurship literature suggests that the gender gap in entrepreneurship results from individual and contextual factors, such as socio-demographic characteristics, entrepreneurial attitudes, family background, and environmental conditions (Klyver et al. 2013; Kelley et al. 2015; Koellinger et al. 2007; Wagner 2007). Specifically, at the individual level, several studies show that women’s entrepreneurial self-efficacy—or the self-assessment of having the required capabilities to start and run a company—is lower than men’s (Koellinger et al. 2013; Wilson et al. 2007). Thébaud (2010) states in her analysis that women are about half as likely as men to believe that they have the knowledge, skills, and experience to start a business. The lower level of entrepreneurial self-efficacy among women is not limited by geographic regions or cultural subsets but is rather a global phenomenon independent of contextual variables (Wilson et al. 2007; Minniti and Nardone 2007; Allen et al. 2007). In addition, studies show that women have stronger fear to fail as entrepreneurs and are thus less willing to take risks (Bönte and Piegeler 2013; Wagner 2007; Caliendo et al. 2009). Several authors also identify women’s restricted access to financial resources as a reason for the gender gap in entrepreneurship (Muravyev et al. 2009; Heilbrunn 2004; Carter et al. 1997; Fairlie and Robb 2009; Marlow and Patton 2005). Furthermore, research reveals that women, compared with men, have fewer and different social networks (Verheul and Thurik 2001; Koellinger et al. 2013). The lower levels of financial and social resources are also important in explaining why new ventures by women suffer from lower survival rates (van der Zwan et al. 2012; Allen et al. 2007).

We suggest that there is a similar gender gap in immigrant entrepreneurship. Several authors emphasize the disadvantages faced by female immigrants in setting up and running businesses, when compared with their male counterparts. Female immigrants usually face greater obstacles to the acquisition of start-up funding, as well as in the development of formal and informal social networks (Vaccarino et al. 2011; Azmat 2013). Furthermore, as in the case of native-born women, immigrant women probably have a negatively biased perception of their own knowledge and skills, show greater risk aversion, and may have less access to the resources required to start and run a business, which may reduce the likelihood of female immigrants to start and run a business (Terjesen et al. 2016; Wilson et al. 2007). Accordingly, we suggest that immigrant women, when compared with their male counterparts, are a priori less likely to launch and run a business, resulting in a gender gap in immigrant entrepreneurship. This means that there is a negative main effect of female gender on immigrant entrepreneurship. Consequently, we formulate our first hypothesis as follows:

Previous research presents evidence that contextual factors influence the gender gap in entrepreneurship. For instance, greater economic development is related to a larger gender gap in entrepreneurial activity (Klyver et al. 2013; van der Zwan et al. 2012; Minniti 2010; Baughn et al. 2006). There is also empirical evidence that cultural and institutional conditions affect the gender gap in entrepreneurship (Baughn et al. 2006; Mueller 2004; Shinnar et al. 2012; Elam and Terjesen 2010; Estrin and Mickiewicz 2011; Klyver et al. 2013). Klyver et al. (2013) present empirical evidence that a higher level of country-level gender equality in the areas of economy, education, politics, and health widens the gender gap in entrepreneurship. Recently, Thébaud (2015) showed that entrepreneurship is a “Plan B” employment strategy for women if there are institutions that facilitate remaining in paid employment. Specifically, Thébaud (2015) documents that women do not need to choose self-employment as a “fallback” employment strategy if they are embedded in institutional arrangements such as paid leave, subsidized childcare, and part-time employment opportunities. Accordingly, the gender gap is greater in more developed economies with supportive institutions.

With respect to female immigrants, however, we argue that entrepreneurship is not a Plan B but rather a “Plan A” employment strategy. In many cases, immigrant women arrive in their host societies with low levels of human capital, inadequate educational qualifications and skills, and poor language proficiency. These factors negatively affect their search for traditional employment (Rath and Kloosterman 2000; Collins and Low 2010). Even when immigrant women possess strong educational backgrounds, sufficient professional qualifications, and adequate language skills, they often find that their skills, competencies, and talents are less recognized and appreciated in their host societies—thereby reducing their chances of acquiring paid employment (Levie 2007; Collins and Low 2010). As a result, entrepreneurship becomes their preferred option—the Plan A for making a living—since it allows them to use their qualifications and experiences in a practical setting to generate income. In addition, we suggest that the marginal utility of supportive environmental conditions is higher for women than for men. Since female immigrants show, compared with their male counterparts, typically lower levels of entrepreneurial self-efficacy and self-confidence, and have less financial and social resources, they may benefit relatively more than men from supportive contexts. Thus, favorable environmental conditions might help female immigrants to overcome the lack of confidence and resources which prevent them from setting up and running a business. Accordingly, this reduces the gender gap in immigrant entrepreneurship. In the following section, we refer to the effects of immigrants’ embeddedness in supportive economic, social, and institutional environmental conditions on the gender gap in immigrant entrepreneurship to develop our specific hypotheses.

We obtained individual and country-level data from multiple sources. First, we used the individual-level data of immigrants from the Global Entrepreneurship Monitor’s (GEM) Adult Population Survey (APS) 2012 database. GEM’s APS provides standardized data on each population’s entrepreneurial preferences, capacities, and activities (Sternberg and Wennekers 2005). The survey is administrated to a representative sample of adults in many countries, making it the most suitable data-driven study of entrepreneurial activity worldwide (Langowitz and Minniti 2007; Urbano and Alvarez 2014). In 2012, GEM added questions to its survey on the subject of immigrant entrepreneurship, allowing the differentiation of individuals according to their country of origin.

Second, we incorporate country-level data from several sources to assess a country’s economic, social, and institutional environment (in which immigrants are embedded). We used GEM’s National Experts’ Survey (NES) 2012 database, which provides background information from selected experts in each country on factors that influence entrepreneurial activity in their respective countries (Xavier et al. 2012). GEM’s NES employs standardized questions and validated measurement scales (De Clercq et al. 2013), developed to overcome the lack of national-harmonized indices or measures that could be used for comparative studies (Amorós and Bosma 2014). GEM’s NES database is widely recognized and has been frequently used in recent entrepreneurship studies (De Clercq et al. 2013; Levie and Autio 2008; Bowen and de Clercq 2008; Martínez-Fierro et al. 2016). In addition, we obtained data from the World Bank database.

To test our hypotheses, we conducted multilevel mixed-effects logistic regression using the “melogit” command in Stata 15. Multilevel modeling is superior to traditional regression techniques because the latter provide inefficient estimates and biased standard errors in the presence of nested data (Snijders and Bosker 2012; De Clercq et al. 2013; Mikucka 2014). In our data set, individuals were nested in countries. Ignoring the interdependency between individual and country-level data may thus lead to artificial significant effects (Snijders and Bosker 2012). To avoid inefficient estimation and biased standard errors, multilevel modeling considers the nested data structure and simultaneously estimates the variability in the dependent variable within and between countries (Snijders and Bosker 2012; Mikucka 2014). Compared with conventional models, the multilevel mixed-effects approach accounts for fixed and random effects to adequately model effects between variables on the different levels of the research design (Snijders and Bosker 2012).

Since we selected only those countries for which all individual and country-level variables considered in this analysis were available, the number of observations was low in some countries. For this reason, we followed the rule proposed for multilevel models and included only countries having at least 30 units at each level of the analysis (e.g., Maas and Hox 2004). Our final sample is comprised of 11,471 immigrants located in 51 countries.

Table 2 provides the correlation matrix. The correlation matrix reveals that immigrant entrepreneurship is negatively linked to gender. Further, we find positive, significant bivariate relationships between immigrant entrepreneurship and education, household income, employment status, start-up skills, know an entrepreneur, household size, and negative and significant bivariate relationships with age and fear of failure.

Table 3 presents the empirical results of our multilevel mixed effects logistic regressions. Before analyzing our models with predictors, we estimated a null model (or intercept only model) for immigrant entrepreneurship. We computed the intraclass correlation coefficient (ICC) which estimates the percentage of total variance in immigrant entrepreneurship that exists between countries. The ICC is calculated directly by dividing the between countries variance by the total variance, and is interpreted as the proportion of total residual variation that is due to differences between countries. We found that about 13.2% of the immigrant entrepreneurship variance occurs between countries. In international business research, intraclass correlation coefficients of 0.05, 0.10, and 0.15 are considered small, medium, and large, respectively (Hox 2010). Thus, multilevel specification is reasonable.

Model 1 includes the control variables, and Model 2 adds gender, the independent variable. Models 3 and 4 include the interactions terms between gender and the economic embeddedness variables. Models 5 and 6 integrate the interaction terms between gender and the social embeddedness variables, and Models 7 and 8 include the interaction terms between gender and the institutional embeddedness variables.

The results of Model 1 show that the control variables age (β = − 0.018; p < 0.01), employment status (β = 1.059; p < 0.01), start-up skills (β = 1.501; p < 0.01), fear of failure (β = − 0.558; p < 0.01), know an entrepreneur (β = 0.891; p < 0.01), and country-level unemployment (β = − 0.030; p < 0.05) are significantly associated with immigrant entrepreneurship.

The results of Model 2 reveal that a gender gap exists in immigrant entrepreneurship (β = − 0.246; p < 0.01). The negative coefficient indicates that female immigrants, compared with their male counterparts, are less likely to engage in entrepreneurship. This pattern is stable. Across all models, a negative relationship between gender and immigrant entrepreneurship was found, indicating that female immigrants are less likely to engage in entrepreneurship. Thus, we confirm Hypothesis 1.

We expected that the gender gap in immigrant entrepreneurship is smaller in countries with supportive environmental conditions. First, we examined the role of immigrants’ economic embeddedness. The results of Models 3 and 4 shown in Table 3 show positive moderating effects of market opportunities (β = 0.507; p < 0.01) and market access (β = 0.457; p < 0.05) on the relationship between gender and immigrant entrepreneurship. This implies that the existence and accessibility of opportunities affects female immigrants more positively. Thus, we confirm Hypotheses 2a and 2b.

Next, we examined the specific effects of immigrant and government networks on the association between gender and immigrant entrepreneurship. The results of Models 5 and 6 provide support for our Hypotheses 3a and 3b. Immigrant network (β = 0.015; p < 0.05) and government network (β = 0.442; p < 0.01) increase the likelihood of female immigrants choosing entrepreneurial activity, compared with male immigrants, thereby reducing the gender gap in immigrant entrepreneurship.

The institutional embeddedness models estimated the moderating effect of business-friendly regulation and rule of law on the relationship between gender and immigrant entrepreneurship. The results of Models 7 and 8 indicate that business-friendly regulation (β = 0.014; p < 0.05) and rule of law (β = 0.008; p < 0.01) positively moderate the gender-immigrant entrepreneurship relation. The gender gap in immigrant entrepreneurship is smaller in countries with better business-friendly regulation and stronger adherence to the rule of law. Thus, Hypotheses 4a and 4b are confirmed.

We also explored whether our findings are robust for forced immigrants. Forced immigrants, who are displaced by violence and conflicts, environmental disasters, or development-induced re-settlement, tend to arrive in their host countries with very limited resources—a fact which makes them relatively more dependent on the host country’s support. Consequently, the relationships that we found for immigrants should also hold for forced immigrants. To analyze the gender gap in forced immigrant entrepreneurship, we considered data of only those immigrants who migrated from fragile states, which typically fail to perform basic functions in the areas of governance, security, education, economic development, and human rights. To define immigrants as forced immigrants, we draw on Fragile States Index data from the Fund for Peace Website. The Fragile States Index assesses the vulnerability of states to collapse based on 12 key political, social and economic indicators and over 100 sub-indicators (Carlsen and Bruggemann 2017). The Fragile States Index sorts countries into four categories, with respect to the country’s fragility and vulnerability: sustainable, stable, warning, and alert. We defined immigrants of the first and second generation as forced immigrants when they (or their parents) migrated from countries ranked as either “warning” (Fragile States Index, 60.0–89.9) or “alert” (Fragile States Index, 90.0–120.0). The results shown in Table 4 reveal that a gender gap in entrepreneurship exists among forced immigrants. In addition, the moderating effects are consistent with those shown in Table 3. Supportive economic, social, and institutional conditions mitigate the negative relationship between forced female immigrant and entrepreneurial activity, thereby narrowing the gender gap in forced immigrant entrepreneurship. A noteworthy observation was that the coefficients of the moderating effects were slightly stronger compared with those shown in Table 3, possibly providing first evidence that the effect of a supportive host country’s environment is even more important for forced female immigrants.

One may assume that the positive moderating effect of the economic, social, and institutional environment on the association between gender and entrepreneurship applies to both native-born and immigrant individuals, implying that our results only mirror a general pattern in the relationship between male and female entrepreneurship. However, we argued earlier that entrepreneurship is a Plan A employment strategy for female immigrants as a result of missing alternatives, whereas it is a Plan B employment strategy for female natives who have a higher level of freedom with regard to job choices. Without differentiating between native-born and immigrant individuals, Thébaud (2015) and Klyver et al. (2013) present general evidence that supportive environmental conditions negatively affect women’s self-employment choice compared with that of men, thereby widening the gender gap in entrepreneurship. According to these findings, entrepreneurship is not the first choice for women when the environmental conditions are better for them. To explore the differences between immigrants and natives concerning the gender gap in entrepreneurship, we conducted the same analysis introduced above for native-born individuals. The results are presented in Table 5. In line with our reasoning, we found significant negative interactions between gender and the contextual factors in five of six moderating models, indicating that good environmental conditions influence the self-employment choices of immigrant and native women differently. In contrast with immigrant women, the gender gap is widened for native women in a supportive environment. Thus, a supportive environment has contrary effects for immigrant and native women regarding their likelihood of engaging in entrepreneurship.

The purpose of this study was to investigate the gender gap in immigrant entrepreneurship—an area that is still largely unexplored. To provide a better understanding of the gender gap in immigrant entrepreneurship, we draw on the mixed embeddedness approach, which emphasizes the central role of the host country’s economic, social, and institutional environment for immigrant entrepreneurship (Kloosterman et al. 1999; Rath and Kloosterman 2000; Kloosterman and Rath 2001; Kloosterman 2003). Previous research documents the important role of supportive and barrier-free contexts for people’s likelihood of starting and running a business (Lüthje and Franke 2003; Pruett et al. 2009; Welter and Smallbone 2011). Accordingly, we argue that context plays an important role in immigrant entrepreneurship. Specifically, we suggest that immigrant women benefit more strongly from a supportive environment than men because women are generally in a more disadvantaged position for entrepreneurship. Previous research documents that female immigrants face greater problems when it comes to start-up financing, the development of formal and informal social networks, and economic rights (e.g., Vaccarino et al. 2011; Azmat 2013), which contributes to their relatively lower entrepreneurial activity. This disadvantaged position should thus result in a stronger need for supportive environmental conditions among immigrant women. If the environment responds to this stronger need of immigrant women, it should affect immigrant women’s readiness to start and run a business relatively more than men’s readiness, as men generally have better access to varied resources and are thus less dependent on environmental conditions that facilitate entrepreneurship.

We tested our hypotheses with data of 11,471 immigrants from 51 countries. Our findings reveal that female immigrants, compared with their male counterparts, are less likely to start and run a business. This result supports the findings of previous literature, which have not been specifically related to immigrants, but have shown that women are less likely to be entrepreneurs than men across countries and cultures (Minniti and Nardone 2007; Koellinger et al. 2013; Klyver et al. 2013; Thébaud 2015). Moreover, we found empirical evidence that the relationship between gender and immigrant entrepreneurship is affected by immigrants’ embeddedness in a supportive economic, social, and institutional environment. The gender gap in immigrant entrepreneurship becomes smaller when the host society provides a supportive environment for entrepreneurship. Our findings indicate that immigrant men, in contrast with female immigrants, are less dependent on contextual factors. Thus, our results reveal that the marginal utility of a supportive context is higher for women than for men. We assume that a supportive national environment shows a relatively stronger positive effect on a female immigrant’s entrepreneurial self-efficacy or self-confidence in one’s own talents, skills, and capabilities. By employing two post hoc analyses, we show that the positive moderating impact of favorable environmental conditions can be replicated for forced immigrants but not for native-born individuals.

This study contributes to the current discourse in immigrant and female entrepreneurship in several ways. First, this study provides new insight concerning gendered patterns in immigrant entrepreneurship from a multi-country perspective. Drawing on a mixed embeddedness perspective that includes an interactionist approach (Frese and Gielnik 2014; Welter 2010), we discuss and show that a favorable environmental context narrows the gender gap in immigrant entrepreneurship. In addition, we show that the moderating effects are robust for forced immigrants who migrated from fragile states. The finding that only the interactions terms (and not the embeddedness variables) are positively and significantly related to immigrant entrepreneurship provides support for an interactionist perspective that individual and contextual factors jointly influence entrepreneurship (Frese and Gielnik 2014; Welter 2010). This means that an interactionist perspective building on the mixed embeddedness approach helps to better understand the gender gap in immigrant entrepreneurship.

Second, we present new evidence that favorable environmental conditions impact immigrant and native-born individuals differently. Whereas favorable environmental conditions mitigate the negative relationship between immigrant women and entrepreneurship, they amplify the gender gap in entrepreneurship among native-born individuals. Therefore, we argue that entrepreneurship is a Plan A employment strategy for female immigrants, whereas it is merely a Plan B employment strategy for native-born women. While recent research argues that institutional support for women may disintegrate women’s self-employment choice (Klyver et al. 2013; Thébaud 2015), this study’s results show that this does not always have to be the case. In contrast with native women, immigrant women benefit from institutional support. Thus, we contribute to the female entrepreneurship literature by providing new insight into how the same environmental conditions affect the gender gap in entrepreneurship differently.

These gender differences in immigrant entrepreneurship decisions offer valuable insights for designing policy measures aimed at fostering female immigrant entrepreneurship. Previous research highlights the positive impact immigrant entrepreneurs can affect for their host societies. By creating jobs, and by contributing to economic growth and innovation (Baycan-Levent and Nijkamp 2009; Marchand and Siegel 2014; Saxenian 2002), immigrant entrepreneurs can help to balance economic development, quality of life, and social cohesion (cf. Levie 2007). Accordingly, policies aimed at economically integrating both immigrant men and women through entrepreneurial activity can foster the welfare of social units, such as families, communities, and societies. Policies to strengthen immigrant entrepreneurship can include various dimensions, ranging from active financial and educational support to formal rules and regulations that guarantee entrepreneurs barrier-free access to markets, low corruption, and strong law enforcement. For instance, business support programs, which are typically provided by national governments and implemented at the local and regional levels, can positively impact financial resources, social networks, and human capital among (potential) immigrant entrepreneurs, thereby facilitating self-employment (Desiderio 2014). Entrepreneurship training, support in administrative issues important for business activity, legal advice on laws and regulations, mentoring, and start-up funding support can be a significant catalyst for entrepreneurial activity among immigrants (Desiderio 2014). Also, structural policies aimed at providing a business-friendly environment—which may include lower taxes for start-ups, labor market flexibility, strong law enforcement laws, and anti-corruption legislation—could positively affect immigrant entrepreneurship, particularly among women as our results show.

Despite the value of our findings, our research inevitably has some limitations. Our analysis is cross-sectional and we thus do not follow individuals over time. We must therefore speak of relationships, and can only speculate as to whether causal mechanisms work behind the relationship we identify. Future research could focus on variations in immigrant entrepreneurship over time as a consequence of changes in environmental conditions; this would offer a better understanding of causes and consequences in immigrant entrepreneurship. Moreover, three of our mixed embeddedness components are based on impressions and judgments of experts, rather than on hard data (Levie and Autio 2008). Since hard data on entrepreneurship issues are not always available for a high number of countries, we assessed GEM’s National Experts’ Survey ratings as the best alternative for our purpose. Future research could consider further contextual variables based on hard data, such as immigrant discrimination (xenophobia, acceptance of certain religious groups), legal issues (fiscal freedom, labor laws, or the economic, political, and social rights of immigrants), social security systems (including child-care), and taxation (e.g., tax treatment of double income) which may explain differences in immigrant entrepreneurship in general and entrepreneurial gender imbalances across countries in particular (van der Zwan et al. 2012; Baycan-Levent and Nijkamp 2009). An interesting further line of research could examine the role of home country-specific factors in immigrant entrepreneurship. An analysis of the home country’s cultural characteristics (social trust, self-expression/emancipative values, uncertainty avoidance, power distance, and entrepreneurial culture) might be helpful for gaining a deeper insight into immigrants’ entrepreneurial attitudes and behavior, gender differences in immigrant entrepreneurial activity, as well as further significant variations between different immigrant groups (Azmat 2013; Basu and Altinay 2002; Clark and Drinkwater 2010; Hechavarría and Reynolds 2009; De Clercq et al. 2013).

The findings of this study must be qualified by the limitations just described. Nonetheless, our results provide first valuable insights into how environmental conditions influence the gender gap in (forced) immigrant entrepreneurship. We hope this study will be a starting point for further (multi-country) research in (female) immigrant entrepreneurship.