1 Introduction

Migrant entrepreneurship can be located at the crossroads of several overarching processes of urban transformation. First, they embody, literally, processes of globalisation whereby ever more regions of the world get connected through multiple linkages. Second, they are providing important goods and services to both firms and private households in urban economies which are not only increasingly geared towards knowledge-intensive activities but also themselves inserted in evermore intricate global webs of interdependence (Kloosterman, Mamadouh, & Terhorst, 2018). Migrant entrepreneurs, in addition, also fit in with the contemporary dominant narrative which stresses the crucial role of entrepreneurship not just for the entrepreneurs themselves but especially for society as a whole through their contribution to economic growth and employment.

In many countries, and notably in (larger) cities, we can observe an increasing number migrant entrepreneurs. This quantitative shift also comprises a significant qualitative change as rising numbers of migrant entrepreneurs are exploring more high-end, knowledge-intensive activities in urban economies. These major developments indeed warrant social-scientific research and accordingly a large body of academic literature on migrant entrepreneurship has emerged. We greatly acknowledge Eduardo Barberis and Giacomo Solano’s call for investigating migrant entrepreneurship from the mixed embeddedness perspective, i.e. using a more comprehensive approach that aims to capture both structure and agency factors. In this contribution, we conclude this Symposium by briefly outlining the development of mixed embeddedness, reflecting on its position within the international field of migrant entrepreneurship research, and by highlighting its key elements.

The archetypal migrant entrepreneur can usually be found in less-affluent residential urban neighbourhoods, where one may come across a wide range of small-scale businesses catering to the everyday needs of the residents: mini-supermarkets, butchers, green grocers, tea houses, haberdashery stores, clothing alterations and repair shops, phone shops, and so forth. The entrepreneurs themselves tend to lack (officially recognised) educational qualifications which funnels them to particular economic activities, which are characterised by relatively low entry barriers. They are often strongly dependent on family and relatives for (often partly informal) labour and also for start-up capital. They typically cater to low-end markets and, hence, mainly compete on price. These migrant entrepreneurs and their small businesses thus offer cheap goods and services while also creating jobs in urban areas where unemployment is frequently substantial (Waldinger, Aldrich, & Ward, 1990; Kloosterman & Rath, 2003; Kloosterman, 2010; Rath, 2000 & 2002).

Close observation, however, reveals that a quite different type of migrant entrepreneur is increasingly present in contemporary urban economies. Instead of being pushed, these migrant entrepreneurs are pulled towards self-employment as they envisage chances for high-growth businesses. These so-called opportunity entrepreneurs (Dheer, 2018) can be found in higher value-added activities such as ICT- development, advanced producer and consumer services, cultural and creative industries, and fast-fashion manufacturing (Saxenian, 2006; Rusinovic, 2006; Brandellero, 2011; Dei Ottati, 2014; Liu, Painter, & Wang, 2014). With the emergence of this type, we can also observe a change in the spatial patterns of migrant-run firms from lower-end shopping streets to more upmarket premises in central business districts, suburban locations and office parks. In addition, these firms are also increasingly present on the Internet. This trend — growth and diversification of migrant entrepreneurship — is nowadays apparent in many advanced economies as documented in the OECD report (2011) Open for Business; Migrant Entrepreneurship in OECD Countries.

From a broader perspective, these different types of migrant entrepreneurs can be seen as embodying two important inter-related global developments. First, they are part and parcel of a profound demographic changes, notably the increase in migration on a global scale that has taken place in the past decades (Lowell, 2007; Castles, de Haas, & Miller, 2013; DESA, 2015). The rise in migration from both poor and rich countries has changed the face of many cities and contributed to the diversification of their populations (Waldinger, 1989; Sassen, 1991/2001). The population of many large cities today has consequently become much more diverse in terms of region of origin, nationality, ethnicity, religion, and economic, social and cultural resources — some observers even speak of “super-diversity” (Vertovec, 2007). This increase in urban diversity along a host of dimensions challenges established forms of identity and belonging, while at the same time providing building blocks for creating new social groupings by connecting people from different parts of the world on the basis of shared experiences, values, attitudes, and consumption patterns (Mansvelt, 2005; Currid-Halkett, 2017).

Secondly, the urban economies in which these migrant-run businesses are located differ from their Fordist predecessors. Advanced or cognitive-cultural urban economies (Scott, 2008 & 2012) are, on the one hand, characterised by knowledge-intensive activities which are often inserted in global production networks and which tend to rely heavily on the use of digital technology: high-tech (e.g. software development), high-finance (e.g. banks), high-concept (legal services) and high-craft activities (e.g. creative and cultural industries). They are, on the other, also home to a broad array of servile workers who provide all kinds of services to the high-end firms and to the highly educated workers (Scott, 2012). Package delivery, cleaning, restaurants and catering, repair and maintenance, yoga lessons and hair dressing are just a few examples of these servile activities.

With the emergence of cognitive-creative economy, an integrated ecosystem with opportunities for entrepreneurs with poor resources as well as entrepreneurs with high resources have come into being. Migrants entrepreneurs can thus be situated at the intersection of demographic and economic changes — two macro trends which come together in many cities not just in the Global North, but increasingly also in cities in the Global East and South (Wang, 2012; Tranos, Gheasi, & Nijkamp, 2015; Dheer, 2018; Kloosterman, 2018).

This mixed embeddedness approach, more than many other approaches, aims at exploring this complex situation. It offers a basic and versatile framework to address key questions regarding the relationship between the resources of migrant entrepreneurs and the kind of opportunities for businesses they exploit. More specifically, the approach allow us to explore on a more structural level the sorting processes of migrant entrepreneurs and the subsequent “ethnic division of entrepreneurial labour.” On a more abstract level, the mixed embeddedness can be seen as a critical realist approach as we ascribe causal power to deeper lying structures — in our case the opportunity structure — which are not directly observable. This way, we aim to shed light on the relationship between agency and structure, a perennial key issue in social sciences.

In what follows we briefly discuss how the mixed embeddedness approach came into being. Thereafter we highlight the basic components of the approach and we argue why such an approach is superior to other, more mono-causal approaches. Our contribution concludes by pointing to the adjustments that the model has undergone since its introduction in the late 1990s.

2 Why the Mixed Embeddedness Approach?

The mixed embeddedness approach is, obviously, not the first attempt to describe, analyse and explain the development of migrant entrepreneurship. We are, hence, building on the work of pioneering researchers. In its initial phase in the 1970s and 1980s, much research was preoccupied with more mono-causal explanations — typically focusing on agency and its resources, for instance by highlighting the role of human, social or cultural capital (Light, 1972; Portes, 1995; Bates, 1997). While they evidently contributed to a better understanding of migrant entrepreneurship, key elements, in our view, were rather absent.

For starters, there was not always substantial empirical evidence to corroborate the more mono-causal explanations. Many scholars, for example, have taken it for granted that migrants set up shop as a response to blocked opportunities on the labour market due to racism and discrimination which will be particularly manifest during an economic downturn when demand for labour in general is relatively limited (Engelen, 2010). In that case, then, we would expect a rise in migrant self-employment in times of economic recession and a fading away in times of economic boom. In reality, however, the opposite is true (at least in the Netherlands). Migrant entrepreneurship generally seems to move in parallel to the economic cycle, just as is the case with mainstream entrepreneurs. Other scholars, following the economic sociological trend to focus on social networks, have emphasized the significance of enclaves or ethnic niches as more or less “safe” spaces for entrepreneurial endeavours (for example Waldinger, 1989). While such an approach could help understand why some entrepreneurs gravitate to particular sectors of the economic and to particular segments therein, they fail to appreciate the significance of changes in the economic and regulatory environment which impact on the demand for goods and services and, therefore, on the opportunity structure for businesses (Rath, 2000). Besides, they understate the sharp competition and many conflicts of interest among those economic actors. In a similar vein, many scholars have argued that entrepreneurial success is contingent on the embeddedness in strong and tight ethnic social networks (Portes, 1995). However, when Turkish entrepreneurs in the Amsterdam garment industry were facing a clampdown on illegal employment practices which made their business ventures practically impossible, their ethnic networks turned out to be useless (Rath, 2002). Accordingly, existing theories on migrant entrepreneurship did not appear to be able to explain a series of concrete empirical manifestations of migrant entrepreneurship.

The empirical shortcomings, however, were also related to more theoretical issues. In particular, we can point to the fact that quite a few studies had taken “ethnic” dimensions of migrant entrepreneurship as their ontological starting point (for example Light, 1972; Light & Gold, 2000; Metcalf, Modood, & Virdee, 1996; Portes, 1987). The authors apparently departed from a view in which ethnic groupings trumped other social divisions. Ethnic groups were thus seen as fundamental basic social units and, consequently, ethnic relations acquired a strategic status. Within such an “ethnic group paradigm”, any activity, notably self-employment, tends to be viewed in a context of ethnic relations and is either part of an assimilationist trajectory or part of a more pluralistic development (which may be the case when the existence and permanence of ethnic niches is emphasized). Economic considerations are pushed aside or to the background. The highlighting of the ethnic dimension in much research on migrant entrepreneurship can largely be traced back to two factors. First, empirical case studies from the United States were dominant in the early phases and they were very much steeped in a tradition of using an ethnic lens. Second, many of the early scholars in this field (notably in Europe) were anthropologists who are more geared towards issues of identity and ethnicity than more general economic developments.

To get to grips with migrant entrepreneurship in a European context, we had to look beyond ethnicity and, moreover, take the broader societal framework into account. As many comparative political economy studies have shown, there are significant differences not just between the United States and European countries, but also among the latter. These differences also contribute to generating divergent opportunity structures thus also affecting the chances of migrant entrepreneurs. Solely focusing on the agency part of the migrant entrepreneurs would obscure these differences and hence leave out an important part of the puzzle. We, therefore, stressed the role of economic and political factors to explain patterns and trajectories of migrant entrepreneurs while relating these to agency thereby adopting an interactionist approach a (Kloosterman, van der Leun, & Rath, 1998 & 1999; Rath, 2000 & 2002; Kloosterman, 2010).

One of the first publications to opt for such an interactionist approach was Ethnic Entrepreneurs: Immigrant Businesses in Industrial Societies. In this jointly authored book, Roger Waldinger, Robin Ward, Howard Aldrich and their American and European colleagues (1990) introduced an interactionist model to analyse “ethnic entrepreneurs” in the United States and Europe. The interactionist model, to quote Craig Mitchell (2015, p. 45), has “acted as a benchmark upon which the mixed embeddedness perspective has been elaborated and developed.” Instead of singling out one particular explanatory factor, they opted for an interactive approach which combined agency and structure. This more comprehensive framework did not just include actors, but also comprised the opportunity structure which they saw as consisting of market conditions (“ethnic consumer products” and “non-ethnic/open markets”) and of access ownership (“business vacancies”, “competition for vacancies” and “government policies”). Although offering a much better and very inspirational framework to grasp migrant entrepreneurship, the approach also attracted criticisms. First, it has been criticised as being too static and more of an instrument for classification than explanation (Rath, 2002). Secondly, the approach still seemed to be based on an a priori categorisation of ethnic or migrant entrepreneurs as being different from their indigenous or mainstream counterparts. Such differences need to be proven rather than taken for granted. Thirdly, the conceptualisation of the opportunity structure is rather limited as it mainly and rather mechanically distinguishes between ethnic and non-ethnic or mainstream markets (Mitchell, 2015). Fourthly, it neglects the wider institutional and regulatory context in which opportunities occur, except perhaps for a small set of rules and regulations specifically for ethnic entrepreneurs.

Notwithstanding these criticisms, this publication was, quite rightly, very influential. Many studies on migrant entrepreneurship have referred to the model introduced by Waldinger and colleagues. Oddly enough, the core message — combining agency and structure — was often ignored in these studies as they tended to focus on one aspect only, namely the mobilization of social networks. “Social embeddedness” obviously matters, but, in our view, this is not enough. The mixed embeddedness approach aims to bring the (opportunity) structure back into the equation by re-introducing the interactive approach and, more importantly, by adding new elements. These elements fall under the heading of three categories: i) divergent opportunity structures on a national or macro-level due to varieties of capitalism; ii) divergent and dynamic opportunity structures on a meso-level due to the emergence of post-industrial or cognitive-cultural urban economies; and iii) the conceptualisation of embeddedness which refers to the key relationships between the micro level of the entrepreneurs and the meso- and macro-level (see also Mitchell, 2015). Below, we will unpack these elements more in detail.

3 Mixed Embeddedness Further Explained

First, the mixed embeddedness approach explicitly acknowledges that opportunity structures are not confined to the market economy per se, but embedded in wider national institutional/regulatory and socio-cultural contexts. Opportunity structures in continental European welfare states are, for instance, notably different from those in Great Britain and the United States. Opportunities for businesses are in essence about markets and national contexts are instrumental in shaping the division of labour regarding allocation through market, state, family or commons (Raworth, 2017) as well as in regulating access to these markets in terms of educational and/or qualifications (Kloosterman & Rath, 2001). Socio-cultural practices partly determine if certain forms of provision are mainly market, family or commons as in for instance child care or care for the elderly. These socio-cultural practices are often at least to some extent formally codified in the national institutional and regulatory framework. Gøsta Esping-Andersen (1990 & 1999) has showed how different national institutional configurations regarding the provision of goods and services have led to different post-industrial employment trajectories. More in particular, relatively high legal minimum wages, and/or family or state provision of all kinds of personal services (e.g. childcare or care for the elderly, food preparation, cleaning and other domestic chores) may hamper the growth of employment in the lower segments of the labour market. In addition, the regulatory framework also affects the opportunity structure, for instance by only allowing entrance to specific markets to entrepreneurs with domestically acquired educational qualifications. One can thus paraphrase Esping-Andersen and state that national institutional contexts will also generate their own specific post-industrial self-employment trajectories and, accordingly, display divergent opportunity structures (Kloosterman et al., 1999; Kloosterman & Rath, 2001; Rath, 2002; Kloosterman, 2010; Jones, Ram, Edwards, Kiselinchev, & Muchenje, 2012; Gertner, Ram, Trehan, & Jones, 2015; Solano, 2016; Ulceluse, 2016).

Institutional and regulatory contexts are typically anchored at the national level. There are, however, also sets of regulations with a more local dimension which can also impact on the opportunity structure as, for instance, zoning plans which may block (or promote) particular businesses (e.g. phone shops or shisha lounges versus creative businesses) altogether or in certain neighbourhoods. It can also be the case that there are differences in local practices regarding the division of labour between market, family and commons. Consumers in Amsterdam on the 1990s were, for instance, much more eager to go to a restaurant than those in Rotterdam thus creating more opportunities in the restaurant and catering sector (Rekers & van Kempen, 2000, pp. 63–64).

Opportunities for businesses in capitalist societies emerge through markets. Markets and all forms of economic action are always and everywhere embedded in larger social and cultural contexts as well as in institutional and regulatory frameworks — this holds for so-called corporatist welfare states such as continental European welfare states but also for liberal welfare states such as the United States just as much just the United States (Granovetter, 1985 & 2005; Esping-Andersen, 1990 & 1999). Labour laws, regulations pertaining to businesses, forms of self-organisation of businesses, traditions of consumer practices and expectations: these are all examples of place- and time-specific contexts which may impact on how markets function and, hence, on opportunities for firms. Entrepreneurs in the United States are, for example, often function in environments in which much regulation stems from civil law instead of public law (e.g. litigation by customers).

Secondly, opportunity structures are more dynamic than Waldinger et al. (1990) assumed. Shifts in opportunity structures can be driven by technological developments and shifts in global trade, but also by changes in regulatory framework (e.g. neoliberal policies shifting provision from state to market) or shifts in socio-cultural practices (e.g. outsourcing of care tasks becoming more accepted). The book Ethnic Entrepreneurs: Immigrant Businesses in Industrial Societies can be seen not only as the start of interactionist approaches to analyse migrant entrepreneurs, but also in retrospect as the end of an era. After its publication in 1990, it became increasingly clear that urban decline had stopped in many cities. Population started to grow gain as young, often highly educated workers (especially single-person and double-income households) opted for urban living thereby changing the social composition of the cities (Musterd & Gritsai, 2009) and, even more relevant in this context, generating demand for all kinds of services (Currid-Halkett, 2017).

In conjunction, after many years of shrinking employment in many cities exemplified by the closure of factories, a turnaround occurred. Deindustrialisation had to a large extent ran its course and service activities now became the engine of urban economic growth. Manufacturing employment had been in structural decline in many western cities in the 1970s and 1980s, and after 1990 the engine of employment growth was to be found in service activities — public services, but more specifically producer and consumer services. This transformation was first observed in cities at the apex of the global urban hierarchy, the so-called “global cities” (Sassen, 1991/2001; Kloosterman, 1996) and, in due time, also in a large number of other cities (Scott, 2012 & 2017). The urban production system thus shifted towards a wide array of mainly service activities with on the one hand, so-called cognitive-cultural activities requiring the input of highly skilled, knowledge workers, and on the other, and a wide range of in-person services or “servile” activities (Scott, 2012). The return of population growth in combination with fundamental transformation of the production system in advanced cities evidently shifted the opportunity structure for businesses and more specifically for small businesses. Whereas Waldinger et al. (1990) observed opportunities for small businesses in mostly stagnant sectors, a whole new set of more promising opportunities opened up with the urban renaissance after 1990. Opportunities thus emerged both for (migrant) entrepreneurs with high and low levels of human capital in these rebounding cities (Kloosterman, 2010 & 2014; Aytar & Rath, 2012).

Thirdly, the mixed embeddedness approach is not just about structures and opportunities, but also about agency and actors and how they are embedded in larger social structures. This part of the mixed embeddedness approach draws heavily on the seminal work of Mark Granovetter (1985 & 2005) who — following in his turn Karl Polanyi (1957) — emphasised how all forms of economic action do not take place in a socio-cultural and institutional vacuum: “… actors do not behave or decide as atoms outside a social context, nor do they adhere slavishly to a script written for them by the particular intersection of social categories that they happen to occupy (…) their attempts at purposive actions are instead embedded in concrete on-going systems of social relations” (Granovetter, 1985, p. 487, & 1992; also Mitchell, 2015, p. 46). Granovetter distinguished between two forms: relational embeddedness and structural embeddedness. Relational embeddedness is closely related to notions of social capital and refers to more personalised relations between different actors. Migrant entrepreneurs are, hence, embedded in a (relatively) concrete networks of social relations with customers, suppliers, banks, competitors, and, often overlooked, law enforcers. In the mixed embeddedness approach, we have adopted the relational embeddedness by stressing the (potential) role of social networks and social capital in explaining trajectories of migrant-run businesses personal, while we have added another, deeper and less directly observable layer of embeddedness by relating actors to opportunity structures and embedding these structures in national institutional and regulatory contexts (Kloosterman, 2010; Rath, 2002). This latter type of embeddedness, then, relates to the broader context and comprises the socio-cultural and institutional settings which help to shape their actions. This concept, hence, transcends direct dyadic, personal relations (Granovetter, 1985).

The mixed embeddedness approach thus links the micro-level of the actors and their resources to the meso-level of the opportunity structures within cities. These opportunity structures, in their turn, are seen as being embedded in the national socio-cultural, institutional and regulatory framework which, on the one hand enable or foster certain markets, while, on the other, block or hamper the emergence and growth of other markets (Kloosterman, 2003 & 2010). This way, another (macro-) layer is added to the interactionist approaches of migrant entrepreneurship research. We thus offer a layered ontological conceptualisation — in line with critical-realist approaches (cf. Ram, Jones, & Villares-Varela, 2017; Jones, Ram, & Villares-Varela, 2018) — with a deeper, not directly observable macro-structure which impacts on the size and shape of the opportunity structure. This extended conceptualisation of the opportunity structure enables a more in-depth and dynamic analysis of the interaction between agency and structures.

In the almost two decades since its launch, mixed embeddedness has become a widely used approach in the field of migrant entrepreneurship research (Aliaga-Isla & Rialp, 2013; Ma, Zhao, Wang, & Lee, 2013; Mitchell, 2015; Dheer, 2018; Jones et al., 2018). It offers a generic and open framework for interpreting the interactions between agency (the entrepreneur) and structure (the opportunity structure). It combines the four analytically distinct key aspects of migrant entrepreneurship as distinguished by Ratan Dheer (2018, pp. 5–8) in his thoughtful analytical overview: i) the opportunities, ii) “enterprising individuals” (the resources of the migrant entrepreneurs), iii) the environment (the wider socio-cultural and institutional context), and iv) “the modes of organising” (business strategies) in systematic but open way. Consequently, the mixed embeddedness approach has been applied to address a wide variety of issues related to the multifaceted phenomenon of migrant entrepreneurship and also to entrepreneurship more generally (e.g. Aytar, 2014; Aytar & Rath, 2012; Cunningham & Tolonen, 2017).

4 What Is New?

The mixed in the mixed embeddedness approach refers first and foremost to integrating the embeddedness of opportunity structures in a broader societal context with the embeddedness of entrepreneurs in social networks in an overarching interactionist dynamic model. It takes into account both the variation in the composition of the migrant population with respect to human capital, social and financial capital as strategic resources as well as the shifts in the opportunity structure which impact on where openings for businesses occur and how they develop over time.

In the model, the resources variable is split in multiple ways between a low (secondary schooling or less) and high level (vocational or academic schooling), but also low and high levels of financial capital, and homogenous and heterogeneous social networks (Kloosterman, 2010). The opportunity structure can be split between structurally stagnating and expanding markets, the latter being linked to the transformation towards a cognitive-cultural urban economy. These different axes and their components may interact in a myriad of ways. A migrant lacking in skills and financial capital can only start a business in a market for which no high financial investments are required and no thresholds in the form of educational requirements exist. These kind of markets are, in principle, open to (nearly) everyone and therefore such markets tend to be overcrowded and dominated by competition on price. By contrast, a highly skilled migrant should be able to access different and more promising markets — also because s/he is probably more bankable and thus able to get access to financial resources. These markets then tend to have relatively high entry barriers lowering the chances for overcrowding and tend to be characterized by competition on quality. There is, obviously, a world of difference between the former and the latter type of migrant entrepreneur — in terms of labour conditions profits, business potential, and chances of survival and success. The partitions between these different types of markets are not watertight and migrant entrepreneurs may move from vacancy-chain opportunities to more attractive markets by acquiring more human, social and/or financial capital, heterogeneous social capital that may enable them to cater to expanding mainstream markets (Mitchell, 2015; Gertner et al., 2015). This is, however, anything but a given trajectory.

The mixed embeddedness approach has developed since its introduction in the late 1990s (Jones et al., 2018). Efforts have been made to refine the approach and to introduce new elements. More attention has been paid to the time dimension. Time in this case may refer to the various stages of the entrepreneurial trajectory. Entrepreneurs may move to other positions on the axes as they accumulate more strategic resources (Kloosterman, 2010; Schutjens, 2014). But time may also refer to political economic changes. Rules and regulations may have changed over time offering new or more market openings, or reducing them. The deregulation programs of the past few decades evidently impacted entrepreneurial opportunities, in the same vain as reregulation or law enforcement does (Rath, 2002).

Recently, researchers have attempted to insert transnational social capital in the mixed embeddedness approach (Gertner et al., 2015; Solano, 2016; Bagwell, 2018; Rath, Solano, & Schutjens, forthcoming). Given the fact that many migrant entrepreneurs are increasingly embedded in social networks which straddle different countries and often different continents such endeavour makes much sense.

In addition, after the prolonged recession in the wake of the credit crisis, the shifts in the opportunity structures have come to be seen in a new light. Markets have, evidently, not responded uniformly to these shocks. Investigating how migrant entrepreneurs have fared after 2008 in different cities in different countries, requiring a longitudinal approach, might also open new insights into the resilience of migrant-run businesses (Riva & Lucchini, 2015).

These timely issues can adequately be addressed by using the mixed embeddedness approach as the point of departure. Its transparent and open design and its inclusion of key variables enables focusing on wide of array of new questions regarding entrepreneurship of migrants and others (Ram et al. 2017; Kloosterman, 2018; Jones et al., 2018; Rath & Schutjens, 2015; Rath et al., forthcoming)