Globalising Swedish countrysides? A relational approach to rural immigrant restaurateurs with refugee backgrounds

ABSTRACT The main purpose of the article is to connect rural immigrants’ business ventures and development in Sweden to relational perspectives on their proximate and distant family and co-ethnic networks at structural and individual levels. Accordingly, the authors employ a relational approach and draw on in-depth interviews. In the context of urban–rural relationships’ meanings for the restaurateurs’ business benefits and constraints, they address two questions: (1) What does embeddedness in proximate and distant family and co-ethnic networks mean for the interviewed restaurateurs and for their businesses? and (2) How do previous and anticipated transitions in the restaurateurs’ families influence their business decisions and migration trajectories? The results suggest that the interviewees employed transnational dimensions in their social embeddedness and that they maintained material and emotional relationships with their countries of origin. This relational approach thus contributes to a better understanding of what the studied businesses mean for the entrepreneurs and the selected localities. The restaurateurs contribute to a globalisation of Swedish countrysides, but their socio-economic potential for countering rural depopulation in Sweden is not fully realised. Additionally, the study illuminates how individuals influence, and are influenced by, place-to-place mobilities on a daily basis and during their life course.


Introduction
Changing rural populations Academic interest in changing rural populations in 'more developed' societies has recently increased (Smith 2007;Woods 2007;Hugo & Morén-Alegret 2008;Bell & Osti 2010;Milbourne & Kitchen 2014;Hedlund et al. 2017). Migrants' motives for migration related to necessity and choice are relevant themes in further studies of rural mobilities and population change (Milbourne 2007;Stockdale & Catney 2014;Findlay et al. 2015). This relates to the fixity and fluidity of globalising rural places (Milbourne & Kitchen 2014), meaning that places of origin and destination are socially constructed and 'never completed' (Thrift 1999), as they consist of intersecting power relations (Massey 1994;Hudson 2004). Woods & McDonagh (2011) argue that globalisation mainly concerns refashioning relations across space on a concrete day-to-day basis. Through the social and economic relations of individuals, groups, and societies, the global and the local reshape each other in negotiation rather than in competition, creating an unequal geography of globalisation.
As a result of increased global interconnectivity, Swedish rural areas are to some extent experiencing repopulation through international migration (Hedberg & Haandrikman 2014). This means that Swedish countrysides previously characterised by relatively homogenous ageing populations and outmigration are currently experiencing the effects of in-migrating younger, international, more heterogeneous populations. However, due to rules, regulations, and strong loyalty to in-group members at the destinations, immigrants living on the margins of Swedish and Nordic societies can experience difficulty in entering or integrating into the labour market (Forsander 2004;Åslund 2005;Kelly & Hedman 2016). Approximately 90% of Sweden's foreign-born working population live and work in cities because finding a job in rural areas may take more time than in urban areas (Salomonsson & Israelsson 2016;Hedlund et al. 2017). This indicates a missed opportunity for minor Swedish settlements, which will become dependent on an inflow of labour (including from abroad) in the near future (Salomonsson & Israelsson 2016).
The majority (38.5%) of Sweden's foreign-born rural population consists of Nordic citizens, followed by East Europeans (15.5%), West Europeans and North Americans (14.4% together), and Asians (13%) (Hedberg & Haandrikman 2014;Lundmark et al. 2014, 430;Hedlund et al. 2017). A minor group (8.1%) is of Middle Eastern and North African origin (Lundmark et al. 2014, 30). The rural tourismrelated sector is interesting to study in this context because people from the Middle East and North Africa show high rates of self-employment in this sector (27.9%), second only to West Europeans and North Americans (41.5% together) (Lundmark et al. 2014, 430). This raises the question of why immigrants from the Middle East and North Africa comprise such a large share of the self-employed in rural tourism in Sweden.
Of all members of the foreign-born population who are self-employed in tourism and living in rural municipalities, 72% are registered as restaurateurs (i.e. they run a restaurant as part of the tourism-related sector) (Lundmark et al. 2014). Compared with, for example, agriculture and forestry, business ventures in the restaurant sector require low investment in terms of rent, maintenance, and social and human capital-related start-up thresholds (e.g. level of education, knowledge of market conditions, or language skills), which facilitates, for example, the pizzeria as an immigrant business (Katila & Wahlbeck 2011;Lundmark et al. 2014). Immigrants who operate small restaurants such as pizzerias contribute to rural restructuring and transnational socio-economic dynamics by creating employment opportunities, diversifying local business structures, and increasing localities' overall attractiveness to tourists and residents (Najib 2008;Hedberg et al. 2012), which can have positive effects in terms of tax income generation, local place image, and the renovation of old buildings (Najib 2008;Stenbacka 2013;Lundmark et al. 2014).

Relational approaches
Relational approaches can shed light on the irregular expansion and acceleration of local, regional, and transnational place-to-place mobilities on a daily basis and over the life course with regard to how individuals influence and are influenced by conditions in multiple places (Thrift 1999;Massey 2005;McIntyre et al. 2006;Benson & O'Reilly 2016). However, few theories or concepts deal with functional and human urban-rural relationships. On a structural level, urban-rural relationships refer to functional linkages and interactions between urban and rural areas (Zonneveld & Stead 2007); they cover a broad spectrum of interactions through housing, employment, education, transport, tourism, and resource use, including social transactions, administrative and service provision, and the movement of people, goods, and capital (Preston 1975;Stead 2002). Associated with these interactions, Nordregio (2000) identifies various types of urban-rural relationships in Europe, of which 'home-work' (regarding employment, accessibility, housing, commuting, and labour markets), 'central place' (local amenities, education, training, commerce, and health and cultural facilities), and 'network' relations (development corridors, conurbations, or polycentricity) are relevant for our study.
Thus, due to increasing urban pressure on their rural hinterland, heterogeneous actors' interest in rural areas is increasing (Overbeek 2006). This means that parts of former rural production landscapes have become landscapes for consumption and recreation by urban dwellers too, 1 although production still remains important in most rural landscapes (Brouder et al. 2015). Therefore, the new rural paradigm implies that policies for rural development must shift focus to include varying economic potential and the needs of different rural areas (OECD 2006). This means that we need to adapt empirical and academic concepts regarding rural areas' mobility, vulnerability, and sustainability patterns (Marsden 2009). It is more plausible to characterise places as part of a continuum from the most rural to the most urban condition, considering differences in population size and density, infrastructure, location, proximity to the built-up area, quality and type of amenities, spheres of influence, cultural distinction, and regional identity (Champion & Hugo 2004;Overbeek 2006;Hedlund 2016). A relevant issue in this context is whether rural areas closer to the urban fringe attract different immigrant groups than do more remote areas (Lundmark et al. 2014, 437). In this article, we examine what urban-rural relationships mean for the interviewed restaurateurs in terms of benefits to and constraints on their businesses. In connecting our empirical data with urban-rural functional linkages (Zonneveld & Stead 2007), social networks (Mitchell 1969), and concepts of embeddedness, based on the work of Granovetter (1985), we address two research questions: 1. What does embeddedness in networks of proximate and distant family and co-ethnic networks mean for the interviewed restaurateurs and their businesses?
2. How do previous and anticipated transitions in the restaurateurs' families, such as marriage, childbirth, divorce, and retirement, influence their business decisions and migration trajectories?
The aim of this study is to connect rural immigrants' business ventures and development in Sweden to relational perspectives on their proximate and distant family and co-ethnic networks at structural and individual levels. Here, proximate and distant networks comprise local and translocal as well as rural and urban dimensions. Hence, the relational approach we take in this article concerns interactions between immigrant restaurateurs in Swedish peri-urban areas and their fluctuating family and co-ethnic networks.

Social networks and embeddedness in urbanrural settings
Globalisation and time-space compression can imply a fragmentation of established territorial patterns of dynamic relations (Giddens 1990). Some scholars state that we need to be more sensitive with respect to socio-spatial and economic contexts, and call for a relational turn in studies of globally interconnected economic practice (Rath & Kloosterman 2000;Fløysand & Jakobsen 2010). In integrating geographical aspects with social and cultural aspects, this turn enables us to understand entrepreneurship among ethnic minorities as based on a mixed embeddedness rather than on affinity with co-ethnic networks only (Rath & Kloosterman 2000). For example, since business owners are embedded within fluctuating networks of social relationships, they take economic and non-economic decisions about their firms depending on their positions and ongoing interactions within these networks (e.g. Shaw 2006). Here, we concur with Aldrich & Cliff (2003, 575), who define entrepreneurship as 'the process by which people discover and exploit new business opportunities, often through the creation of new business ventures'. This definition is relevant, since our study concerns how immigrant entrepreneurs are embedded in local and extra-local social spaces, and how they may exploit business opportunities using business and other networks across different geographies (i.e. urban and rural areas in Sweden and in their countries of origin) (Young 2010).

Social network theory and embeddedness
There is limited knowledge of immigrant entrepreneurs with refugee backgrounds, and the meaning of their co-ethnic and family networks for rural entrepreneurial endeavours, but a number of studies similar to ours used social networks and embeddedness as analytical tools (Neergaard et al. 2005;Bagwell 2008;Wang & Altinay 2010;Munkejord 2017). Mitchell (1969) developed social network theory to designate how people understand the implications of their involvement and position in certain social networks and relationships for their social action and behaviour. In search of relatively cheap resources, information, and other emerging business needs, entrepreneurs can activate different social networks, depending on their size, positioning, and relationship structure (Greve & Salaff 2003;Bagwell 2008). This results in multiplex strong and weak ties with several layers and different content, which play myriad roles in an entrepreneur's social context. Strong ties can exist with reliable people such as friends and family who hold similar views and beliefs, and use problemsolving techniques, whereas weak ties can exist with people who offer useful information but otherwise operate outside an individual's immediate network (Granovetter 1973;1995). However, diasporic and weak ties may also provide access to information and social networks that may prove more important than opportunities offered through strong ties (Putnam 2000).
The line of reasoning relating to the roles of strong and weak ties in social networks constitutes the basis of Granovetter's (1985) theory of embeddedness, which asserts that entrepreneurial decisions regarding business ventures and development do not emerge in a void but are rather embedded in concrete ongoing systems of social relations (Granovetter 1985). Connecting entrepreneurship with family embeddedness provides a frame for studying immigrant entrepreneurs' unique cultural, financial, human, and physical resources (Aldrich & Cliff 2003). More specifically, this set of family resources consists of family members' possible willingness to listen to new business ideas and offer advice, and to provide both intangibles (access to information, networks, knowledge, and support) and tangibles (financial capital and unpaid family labour). Thus, the family may embody economic, cultural, and social obligations, and investments stemming from before migration and continuing after migration during its adaptation to the host society (Sanders & Nee 1996). It follows that family transitions such as marriage, childbirth, divorce, and retirement are important contributing or constraining factors in business processes and achievements (Aldrich & Cliff 2003). It has been indicated that family embeddedness is crucial for ethnic minority entrepreneurs to start and run a fledgling business successfully (Rath & Kloosterman 2000;Bird & Wennberg 2016).
Co-ethnic networks can imply similar dynamics, consisting of co-ethnic suppliers, customers, and business colleagues who speak the same language. Among them, a migrant entrepreneur can get business advice or recruit co-ethnic workers who are willing to work for a modest wage (Wang & Altinay 2010;Andersson Joona & Wadensjö 2012). For example, some Turkish staff accept working under poor conditions in kebab shops because they value collaboration with members of their co-ethnic networks and see this as a way into the labour market (Wahlbeck 2007). It is likely that members of co-ethnic networks play a role before, during, and after migration, since they are present in both the areas of origin and destination. Therefore, such networks are important in the selection of a preferred destination, as well as the adaptation and acclimatisation there (Drori et al. 2009). Both family and co-ethnic networks can contribute to linking migrants, former migrants, and non-migrants in communities in both places of origin and places of destination (Massey 1988). Hence, although people other than those in migrant entrepreneurs' own ethnic groups and extended families can be considered reliable, trust and social capital are significant for the existence of an ethnic employment pattern, whereby co-ethnic networks in the restaurant sector can assist in providing jobs (Putnam 1993;Bagwell 2008;Munkejord 2017).
Previous studies of proximate and distant family networks' implications for migrant entrepreneurship have shown different dynamics (e.g. Munkejord 2017; Webster 2017; Webster & Haandrikman 2017). On the one hand, geographical proximity is an important criterion for capitalising on family resources. Being close to reliable persons such as family members improves the recognition of business opportunities, communication around economic decisions, and interpersonal cooperation and dialogue (Bird & Wennberg 2016). On the other hand, developments in technology and mobility have facilitated more dispersed social relations. Immigrant families' transnational ties therefore play a major role by providing ideas or identifying workers with the necessary skills, who will accept the often harsh working conditions in immigrant businesses (Munkejord 2017). Hence, transnational family networks consist of various types of sustained family relations and activities conducted across national borders, and embody spatially enlarged sources of family embeddedness that are not usually available to mainstream entrepreneurs (Glick-Schiller et al. 1995;Vertovec 2003;. Although the theory of embeddedness is criticised for being stretched to encompass almost everything, to the extent that it is meaningless (Krippner et al. 2004, 113), entrepreneurship scholars generally view it as a valuable frame. As examples, 'geographic embeddedness' (Fløysand & Sjöholt 2007) has been used to study an individual's perceptions of entrepreneurial opportunities in a locality, based on its geographic location, demography, and particular assets, while 'institutional embeddedness' (Welter 2011) has been used to study the effects of laws, regulations, and migration policies in host and home contexts on market entry. Additionally, firm owners' 'business embeddedness' in rural and other areas has been analysed through their local and translocal networks and activities (Dubois 2016). Thus, our study contributes to filling the gap in research on immigrant entrepreneurs with refugee backgrounds and their social networks in connection with functional linkages between urban and rural areas (Zonneveld & Stead 2007), as well as institutions, rules, and regulations (Krippner et al. 2004, 114). Similarly, the majority of interviewees in our study identified themselves as Iranians or Kurds (originating from Syria, Turkey, and Iraq). According to Aradhya et al. (2017), Kurds and Iranians constitute c.7.5% of all immigrants in Sweden. Labour migrants, mainly from rural Turkey, migrated to Sweden during the 1960s and 1970s, followed by mainly Kurd refugees and family reunification migrants, beginning in the 1980s (Westin 2003;Aradhya et al. 2017). Specific employment statistics per origin country are not publicly available, but in 2010 the employment rate among native Swedes in urban areas was 81% and in peri-urban areas 80% (Hedlund et al. 2017). Among Asian immigrants, the employment rate was 49% in urban areas and 31% in peri-urban areas, and among people from Africa and the Middle East it was 46% and 31% respectively (Hedlund et al. 2017). In 2016, 1415 Iraqi, 678 Iranians, and 657 Turks migrated from Sweden (Statistics Sweden n.d.). Syrians were not among the top 20 nationalities emigrating from Sweden in 2016.

Iranians and Kurds in Sweden
In the 1980s and early 1990s, more than 50,000 Iranian refugees migrated to Sweden, where they faced challenges trying to secure meaningful work (Kelly & Hedman 2016). In Sweden, the urban concentration of refugees in general and Iranians specifically has been studied in the context of residential segregation and inter-urban or intra-urban residential relocation (Borgegård et al. 1998;Andersson & Solid 2003). However, relocation to areas that are less urban has been associated with greater degrees of assimilation into host societies (Hedberg & Haandrikman 2014;Webster 2017), which is one important reason why we chose to investigate the role of rural pizzeria owners' distant and proximate networks. Partly due to relatively high unemployment, c.125-250 Iranians returned to Iran annually during the first decade of the 21st century, and 375-750 highly educated Iranians left Sweden annually in search of labour market opportunities elsewhere, such as London or Toronto (Kelly & Hedman 2016).
In Sweden, the 'all-Sweden strategy' partly explains former refugees' current residential patterns (Borgegård et al. 1998;Aradhya et al. 2017). Between 1985 and 1991, the policy aimed to relocate refugees randomly across Swedish municipalities in order to alleviate local administrative and economic pressures and to facilitate integration (Borevi 2012;Aradhya et al. 2017). It has been debated whether the 'all-Sweden strategy' has actually contributed to increasing levels of integration, and whether a longer duration of stay in Sweden leads refugees to remain in the rural places they were assigned to (Borgegård et al. 1998;Åslund 2005;Aradhya et al. 2017). Although the Government of Sweden intended to disperse refugees, the urban concentration of these populations has remained high (Andersson et al. 2010).
Since the abolishment of the refugee placement policy in the 1990s, refugees in Sweden have been free to choose any place of residence after receiving permission to stay (in contrast to, for example, Norway, where they are assigned a location by the Norwegian Directorate of Immigration, UDI). Åslund (2005) shows that newly arrived immigrants are generally more mobile than natives, and prefer larger urban centres to rural areas as residential locations. This is explained by high representation from the immigrants' country of birth and generally large immigrant populations in urban areas, and by the fact that the availability of local public services and access to local labour markets in rural areas is limited. After initially residing in sparsely populated locations, former refugees often move to larger municipalities (Åslund 2005;Aradhya et al. 2017).
However, an increased concentration of immigrants and their businesses in Swedish urban areas has led to tough competition in the restaurant sector (Najib 2008;Lundmark et al. 2014), and resulted in immigrant business start-ups in rural areas. Hence, many immigrants with refugee backgrounds include both urban and rural places in their social and business networks after arriving in Sweden.

Research areas and data gathering
In 2015, we visited Swedish-style rural pizzerias, which resembled kebab and similar fast-food restaurants in other Nordic countries (e.g. Wahlbeck 2007; Katila & Wahlbeck 2011), rather than traditional urban Italianstyle pizzerias. We made two broad selections, one regarding Sweden's rural geography and the other regarding interview subjects. First, we selected localities in the counties of Värmland and West Bothnia (both bordering Norway), since both counties consist of a medium-sized regional capital and a variety of adjacent rural areas (Hedberg & Haandrikman 2014, 131). These geographical settings offer a plethora of urbanrural relationships in both a central part (Värmland) and a northern (West Bothnia) part of Sweden, which are affected differently by globalisation. West Bothnia was selected because of the prominent presence of a regional centre, Umeå (population ≈ 120,000) (Statistics Sweden n.d.), while Värmland was selected because it has a similar regional centre (Karlstad, population ≈ 88,000), which is surrounded by rural spaces and larger municipalities with small towns (Ackerby 2010) (Fig. 1). We also selected the latter county to compare our results with a previous study of another in-migrant population there (Eimermann 2016). Second, we targeted pizzeria restaurateurs for an in-depth interview study. This selection was based on previous research (Najib 2008;Stenbacka 2013) and lay discourse, according to which many rural pizzeria restaurateurs in Sweden were from the Middle East and North Africa. Apart from Najib (2013) and Webster (2017), few in-depth interview studies specifically targeting non-EU restaurateurs have previously been conducted in the two counties or in similar counties in Sweden.
We approached 30 pizzerias in rural localities within commuting distance of urban centres, whose owners could live either in a rural or in an urban location. Hedberg & Haandrikman (2014, 131) propose a rural typology consisting of four categories, based on the distance within which 20,000 people live: adjacent to urban, accessible, peripheral, and remote countryside. Since our study focused on urban-rural dynamics and globalising countrysides, we visited pizzerias located in the categories 'adjacent to urban' and 'accessible' (Fig. 1). Thus, peripheral and remote countryside (Hedberg & Haandrikman 2014;Hedlund 2016) were outside the scope of our study.
In West Bothnia, we approached 16 restaurateurs, of whom 7 were interviewed, while in Värmland we approached 14 restaurateurs, of whom 5 were interviewed (i.e. 18 restaurateurs declined to participate in the study). All interviews were conducted by the authors in Swedish, since most interviewees were fairly fluent in the language; one relatively newly arrived restaurateur used his smartphone for assistance in translation. The interviews were conducted outside peak hours, mainly between 09.00 and 10.30 or between 14.00 and 16.00 on less labour-intensive days of the week. Others present at the times when the interviews were held were often staff with little proficiency in the Swedish language. Hence, although they might have influenced the views expressed during the interviews, we could reasonably assume that this was seldom. We communicated in a rather informal style, to diminish possible power inequalities during the interview situation (Rose 1997). As a consequence, we were able to obtain the interviewees' informed consent and conducted 12 semi-structured individual interviews lasting 30-60 minutes with pizzeria operators of Middle Eastern and North African origin at their restaurant, where they felt at ease.
We based the interviews on a guide with four parts. The introduction regarded geographic, socio-economic, and demographic backgrounds. The second part considered the migrants' social experiences both before and after migration, and the local effects of their presence in Sweden. As an example, one question in this part was: 'Why have you migrated to Sweden?' The third part considered socio-economic aspects in relation to the interviewees' assessments of opportunities, necessities, and combinations of these aspects, by asking, for example, 'Why did you start or acquire this business?' The fourth part considered the interviewees' expectations with regard to their work and living environment in the coming three years. Due to the sensitivity of some information, 10 of 12 interviewees preferred not to be tape-recorded while being interviewed. In these cases, one of us conducted the interview while the other took detailed notes.
Directly after each interview, we completed our detailed notes by elaborating them in the words of the interviewees. We then analysed them thematically (Kvale 2007), using a word-processing program. The results are reported below, and pseudonyms are used to preserve anonymity.

Methodological challenges
The main methodological challenge was approaching and contacting the restaurateurs. We located pizzerias via the Internet and during field trips. We also received information on additional pizzerias by using different starting points for snowball sampling (Hennink et al. 2011). An interview was often preceded by several visits.
When neither telephone calls nor initial e-mails resulted in a response, we visited the selected pizzeria, where we were often met with a friendly response of 'Sorry, but the owner isn't here today.' As noted by other researchers (e.g. Najib 2013), it was often difficult to arrange a meeting for an interview, or the information we received was limited, which is why 18 potential interviewees were not included in this study.
Although we do not include empirical data on the restaurateurs who declined, we briefly reflect here on a number of possible underlying reasons why they did not wish to participate. First, language skills were often insufficient, and some restaurateurs did not understand why we wanted to interview them, were sceptical of our intentions, or declined because they were worried about the possibility of not understanding our questions. Second, the prospective interviewees may have been reluctant to participate due to previous unpleasant experiences with state officials or other interrogators, either in Sweden or in their country of origin. Third, since the restaurateurs worked hard, it is possible that they could not find time to be interviewed.
Hence, there was a bias in the interview study. In many cases, those who agreed to participate had been in Sweden several years and had improved their language skills and knowledge of Swedish culture and habits (Slavnic 2004). They were also less reluctant to be interviewed by a representative of a state-owned Swedish university. Hence, some of the answers should be treated with caution; for example, that stated numbers of staff might have differed from the actual numbers. Nevertheless, valuable information was gathered on the role of family and coethnic networks in the interviewees' migration histories, their motives for migrating and for starting a rural pizzeria, their previous occupations, and their work and migration-related prospects (Kvale 2007).
Contra our intentions, all interviewed restaurateurs were males. This indicates gender issues, which would be interesting to examine in future studies of similar restaurants in the context of migration and rural labour markets. Although we did not interview spouses, we include them in Table 1 because the male restaurateurs referred to their importance in migration and business decisions. Where possible, also the spouses' education and previous work experience are indicated in Table 1, in order of migration, followed by their geographic and socio-economic backgrounds as indicated by the interviewees, for whom we have used pseudonyms. Since some had migrated due to direct threats from current war situations in Syria (e.g. Saud), and others had migrated due to uncertain political circumstances in Turkey (Isam) and Iran (Farat) during the 1980s and 1990s, it was hard to determine who was a refugee and who was not. Still, based on their motives and since the interviewees came from countries associated with refugee migration (Aradhya et al. 2017), we refer to them as restaurateurs with refugee backgrounds, implying that they migrated either as refugees or for family reunification. Since the migrants had arrived in Sweden between 1984 and 2013, we investigated what they did for a living after arrival, which was between 2 and 31 years.

Results
Since all studied restaurants were located in rural areas, we assumed that the owners lived in rural areas. Many of the study interviewees had lived in smaller rural places directly after arriving in Sweden, some as part of the 'all-Sweden' strategy. Others had moved to rural Sweden to be near family members, such as Tomas, who explained that he had done so because his sister had previously migrated there. However, all interviewees had moved from their initial location, and seven of them preferred commuting to their rural workplace while living in an urban centre. Amed said: 'I'd like to live in the same community as where my restaurant is, since it's a tranquil village with easy access to nature.' However, his wife preferred living in a city due to the availability of facilities and employment opportunities, education opportunities for both adults and children, and a larger co-ethnic network. This line of reasoning, which centred on trust and family and co-ethnic ties, was often stated by the interviewees.
The majority of the commuters had previously worked in restaurants in an urban centre. However, they became self-employed restaurateurs outside the urban area due to high competition and high rents in the cities; Izmi offered an example: 'In the city, you'd need to develop a special niche to stick out, perhaps serving Turkish specialties.' For similar reasons, two interviewees had moved from the urban centre to the adjacent countryside in order to take over the running of existing pizzerias. Husain explained: 'In the city, we lived in an apartment, since we couldn't afford a house. The banks wouldn't give us a loan. But here, after running this restaurant for some years, we've been able to buy a house!' Thus, Husain's choice of residence was related to income optimisation and more affordable housing in rural areas. Two restaurateurs in Kristinehamn lived in more urban settings than the settings of their restaurants. As a municipality in a densely populated region, Kristinehamn thus offered sufficient amenities for the restaurateurs and their families.
The information gained from the small study sample suggests that rural pizzeria restaurateurs would rather  ♂ 1962, ♀ unknown (1988, 1997, 2002) ♂ 1 year university studies, economy ♀ university (medicine) ♂ bank clerk (Iran), taxi driver and restaurants^(Sweden) ♀ job in town 1998 Boutal City (population ≈ 2 million), Iran ♂ 1963, ♀ 1978 (2000, 2003) ♂ secondary school ♀ secondary school  1978, ♀ 1978 (2007, 2012, 2013) ♂ low ♀ no formal education ♂ construction* (Syria), restaurants^(Sweden) ♀ no work experience Notes: ** = pseudonym; * = self-employed;^= employed in tourism live in urban areas, provided that they are within commuting distance and the infrastructure is sufficient. If the population in the restaurant's location is relatively large and includes co-ethnic groups, and the level of amenities is developed accordingly, this may imply that restaurateurs are inclined to live in such municipalities. Thus, although the studied rural areas adjacent to and accessible from urban centres offered potential for operating pizzerias, few restaurateurs lived in them.

Immigrant restaurateurs embedded in social networks
Four interviewees originated from Iran and had arrived in Sweden in the 1980s and 1990s, while six interviewees from Kurdistan had arrived in 1999 or later (Table 1). Other interviewees originated from Algeria and Turkey (not Kurdistan). A comparison of population sizes in their Swedish living environments with those in their areas of origin, as estimated by the interviewees, indicates the urban-to-rural direction of about half of the migrants.
Most interviewees were in the age range 18-40 years when they migrated. They had moved with at least one family member, and their motives for choosing Sweden were related to trust, family ties, and co-ethnic networks. They had friends or relatives living in Sweden, who could help them find a place to live, provide contacts, or introduce them to local habits and culture. Some interviewees had social networks in other European countries, and had been advised to choose Sweden because the country has a good reputation regarding human rights, with high degrees of security, democracy, and individual freedom. Husain said: 'Sweden is the opposite of my home country, which is exactly what I was looking for.' Determining the interviewees' exact level of education was difficult, as they were not always able to indicate the duration or degree of their education due to differences between the systems in Sweden and their home countries. However, of the 12 restaurateurs and 8 spouses whose education is specified in Table 1, 4 restaurateurs and 4 spouses specified a high degree of education: they had either studied at a higher education institution for at least one year or had obtained professional training, and 3 of them had had high-skilled jobs, respectively as a teacher, bank clerk, and a business administrator. Hence, becoming a rural restaurateur implied deskilling for 4 of 12 interviewed restaurateurs.
Most interviewees had been employed in restaurants before becoming self-employed restaurateurs. Six had worked in the restaurant sector before migrating, and an additional four had worked in the same sector in Sweden before acquiring a restaurant. Including the two who had sold ice cream or jackets to tourists in their countries of origin, 10 interviewees had previously worked in tourism. Four had been self-employed before operating their current business. Almost all of the interviewees indicated that operating a pizzeria required working up to 80 hours a week, especially in the start-up phase.
Better prospects for their children's upbringing played a major role in the restaurateurs' rural business ventures and choice of residence location. The interviewees' main motives for becoming restaurateurs can be categorised in two groups. Four interviewees saw their business as an opportunity, claiming they had made their choice to run pizzerias as a way to invest in their quality of life in the future. By contrast, eight interviewees saw their business as more or less a necessity, due to lack of education combined with inadequate language skills and limited knowledge of the local market and culture. Although there were distinctions between their narratives, most interviewees saw both opportunities and necessities in becoming self-employed restaurateurs.
The interviewed restaurateurs described were of the opinion that necessities had 'pushed' them into the rural restaurant business. Farat said: 'When we migrated to Sweden, I never thought I'd run a pizzeria! I thought I'd continue my studies, and my wife also planned to study medicine, and to work as a doctor.' Similarly, Amed clarified why he had become a restaurateur, illustrating a rather common trajectory related to trust and social ties, which emerged in several interviews: I'm no good at studying, but I have experience selling ice cream in my home country. In Sweden, I've worked in restaurants, like the one owned by my father-in-law. When you've been employed at a restaurant, you've learned how to do things and you can work there even without the owner being around. Then, you can become a co-owner or buy the place. (Amed, restaurateur, 2015) Thus, the majority of the interviewees confirmed that owning a pizzeria was one of few ways to avoid unemployment. With a sour smile, Brahi compared working at his restaurant to a nightmare rather than a dream: 'I have no education. I was looking for a job and came to be employed here. Later, I bought this restaurant, since it's better than being unemployed, and if someone else had bought it, I could've lost my job. It isn't easy.' The quote is representative of the necessity recount by most of the interviewees.
By contrast, some restaurateurs indicated material and emotional opportunities. Husain stated: 'In my job as an assistant nurse I would never have been able to earn enough money to buy a house, but now I am!' Onur claimed: 'I see this purely as an investment. The location of this restaurant is good, in a locality with an increasing population, so in a few years I should be able to return to my previous job as a consultant, earning extra money from this restaurant.' Isam explained: 'This restaurant is one of several restaurants I own in Sweden. I'm an entrepreneur, you know, and I still own a business in Turkey as well!' Many of the interviewees either owned or had owned several restaurants simultaneously. Many had been motivated by the prospect of buying, selling, or owning a particular restaurant, thus indicating the importance of opportunities to increase their respective family's quality of life. Such individuals were thus active agents, who had taken initiative despite their relatively deprived situation.
Whether perceived as opportunity or necessity, or both, the restaurants contributed to job creation in rural Sweden. Table 2 shows that the 12 interviewees were working full-time, and had created eight parttime and two full-time jobs for their relatives. They indicated 25 additional staff members (11 full-time and 14 part-time), of whom 14 were in their co-ethnic network.

Immigrant restaurateurs' family transitions over the life course
Most of the restaurateurs planned to live and work in more urban areas. The older ones were considering retiring, while the younger ones were pondering their future business and quality of life. Isam had big plans: 'I'm always planning. I'd like to demolish the pizzeria and develop the large plot of land it's on [and] perhaps build a youth hostel or cabins there.' Husain thought of returning to his previous job as an assistant nurse in the city, which would allow him to spend more time with his family: 'I don't think I'll be able to run this restaurant much longer. It would be impossible, for instance, to combine it with activities for the children, such as when they start playing sports'. Onur explained why he had previously claimed to see his restaurant as an investment as follows: From the moment the accountants made their calculations, I knew it would be possible to make a profit. It's all part of a project. To invest money and build up something. This place was a restaurant before, but it was nearly bankrupt. So it was cheap to buy, not so expensive to renovate, and later perhaps I can sell it for a good price. Similar to buying, refurbishing, and selling an apartment. (Onur, restaurateur, 2015) Furthermore, Onur delivered food to a local elderly women's home in order to improve his reputation, and asked non-local musicians in his co-ethnic network to perform during music evenings. Other restaurateurs had similarly made investments, for instance by creating meeting places and other activities for local populations. Husain proudly showed us the award he had recently won for good restaurant service. Amed deliberately planned social activities on various evenings for different neighbourhoods, with the aim to 'make neighbours meet' at his restaurant. Rotan indicated he would renovate and expand his restaurant, having recently organised a series of successful music evenings as a popular local alternative to spending an evening out in the city. Similarly, a member of Brahi's co-ethnic network, who operated a camping site elsewhere, had renovated Brahi's restaurant. Tomas's pizzeria had become a meeting place, but there was little integration between native locals and his Kurdish co-ethnic network. Thus, the restaurateurs and their businesses were embedded in family and coethnic networks. All interviewees except Onur were opposed to the idea of having their children work at their restaurant. Boutal exclaimed: 'Never in my life will I let my children work here!' His reason related to the necessity for long working days, little leisure time, hard work, and years of patience to improve the next generation's prospects in life. Saud said: 'It's like a computer game […] You finish one level before starting the next, and after many levels you've improved your business, and your life.' However, the tough working conditions also led many restaurateurs to think of other options than being self-employed at a restaurant. Brahi mentioned his plans to move to a city in another county, where his relatives lived. Together with a cousin, he wanted start a small car dealership but, first, his cousin had to finish his education and improve his economy. Brahi was not the only interviewee who was seriously considering a move from the countryside to a city. Many interviewees with children over the age of 17 years indicated that their children were studying and aiming for other careers. Farat and Boutal, who were strongly embedded in their family networks, both indicated that their whole family might follow their children if they chose to move to Stockholm for higher education: One of our daughters may start studying in Stockholm. Strong family ties are important for us, so in that case the rest of the family are considering going with her. I may lease out the restaurant for a while and try to find a job in Stockholm. If it doesn't work, I can return here, stop the lease and run the restaurant again. (Farat, restaurateur, 2015) Even more prominent were strategies of combining their current place of residence with living in their country of origin or in a warm place during the winter. At the time when the interviews were held, five interviewees' social ties provided them with access to an apartment or a house in their country of origin or a country with a warm climate. Isam owned his own house in Turkey, where he lived when running his business there five or six times per year. Izmi and his Swedish wife were seriously considering living in Sweden during the summer and in Turkey during the winter when they retired. Additionally, participants who did not own a house in their country of origin had thought about regular return visits to see their relatives.

Discussion
The aim of the study on which this article is based was to connect rural immigrants' business ventures and their development in Sweden with relational perspectives on their proximate and distant family and co-ethnic networks at structural and individual levels. Due to the small sample, it is not possible to discuss these interviewees in separate groups based on their national and ethnic backgrounds. However, some overall key findings can be identified, which complement previous broader quantitative analyses relating to rural restaurants (e.g. Lundmark et al. 2014) and refugees' post-migration residential decisions in Sweden (e.g. Aradhya et al. 2017).
Our findings revealed what urban-rural relationships meant for the interviewed restaurateurs, in terms of benefits to and constraints on their businesses. On the meso-level of social networks, we identified several stages in the business ventures and their development. Many interviewees had previously worked in restaurants or tourism, both in Sweden and elsewhere. Some had migrated from urban touristic areas in their countries of origin to less touristic Swedish rural areas. After arriving in Sweden, all interviewees had moved between various urban and rural places (Champion & Hugo 2004, Overbeek 2006Hedlund 2016). Thus, whereas Lundmark et al. (2014) assumed that restaurateurs who operate in rural areas also live there, the results of our study indicate that although their businesses may be located in peri-urban areas, the restaurateurs often live in cities. Their rural business ventures were motivated by high rents and rising competition in the cities, which led to poorer possibilities for such business ventures there. However, these constraints were less important than the socio-economic benefits offered by proximate and distant family and co-ethnic ties, such as for Isam, who both owned several pizzerias in Sweden and regularly travelled to his business in Turkey. These novel insights regarding transnational practices constitute one key finding, which reiterates and adds to findings from previous studies of the importance of the role of trust and strong ties for migrant entrepreneurs (Sanders & Nee 1996;Drori et al. 2009 Furthermore, the benefits offered by proximate cities consisted of the possibility to commute between the rural restaurant and the city, with its local amenities such as education for children and commerce (Nordregio 2000), as well as proximate family and co-ethnic ties. Such cities thus provide cultural facilities for family members and members of their co-ethnic networks who are not employed full-time, but who offer intangibles and tangibles when needed (Aldrich & Cliff 2003). With respect to Nordregio's types 'home-work' and 'central place', these urban-rural relationships are relevant for the studied restaurateurs (Zonneveld & Stead 2007, 442).
'Home' was often in urban areas, while 'work' was in adjacent and accessible rural places. The towns of Karlstad and Umeå are central places in Värmland and West Bothnia, respectively, to which many people commute from the surrounding hinterlands (Holm et al. 2013). By contrast, most of our interviewees commuted in the opposite direction: they worked in the countryside and mainly lived in the city, with other members of their household. Moreover, those who lived in rural areas were seriously considering moving to a city. However, as illustrated by Farat, who was considering leasing his restaurant (not selling it), the migrants' dynamic relations with myriad places did not mean that their social ties had become separated from their local context (Giddens 1990). Future studies might examine these patterns more thoroughly using register data, and include gender issues and household members' workplace locations (Neergaard et al. 2005).
However, migrant restaurateurs' potential in terms of countering rural depopulation in Sweden has not been fully realised. Our qualitative study shed new light on the complexity of rural-urban networks (Hedberg and Haandrikman 2014;Hedlund et al. 2017). The interviewees' socio-economic strategies in terms of local labour and consumer markets, and employing co-ethnic staff with the required knowledge, skills, and competences, contributed to a flexibly increasing heterogeneity of actors in repopulating Swedish countrysides (Overbeek 2006;Najib 2008;Hedberg & Haandrikman 2014;Webster & Haandrikman 2017).
We call for increased attention to temporalities and more longitudinal qualitative studies of migrant entrepreneurs' businesses. However, it is hard to distinguish between refugees, economic immigrants, and immigrants with refugee backgrounds (Forsander 2004). When does a refugee or an immigrant become a 'normal' citizen? In this article, we have referred to our interviewees as immigrants with refugee backgrounds, but some of them had lived in Sweden for 20 years or more, and the threat they had once fled from has since diminished or disappeared. The interviewees' accounts of flexible attitudes and plans for transnational mobility indicated potential transformations in their migrant status. Hence, future research should investigate how such transformations are shaped over the life course and in relation to developments in the country of origin. For example, considerations of former refugees' ongoing quality-of-life could resonate with understandings of lifestyle in migration (Benson & O'Reilly 2016) and lifestyle-led multiple dwellings (McIntyre et al. 2006). This would have important implications for debates relating to national refugee placement policies, such as the all-Sweden strategy, since refugees and family reunification migrants may be more mobile and incorporate more transnational ties than they appear to be or do.

Conclusions
Addressing the first research question, we find that the interviewees and their businesses employed transnational dimensions in their embeddedness in proximate and distant family and co-ethnic networks. The restaurateurs' social networks consisted of various types of sustained relations with family members and people with the same or similar ethnic background in their areas of origin and destination, and elsewhere. These spatially enlarged sources of family embeddedness played a role before, during, and after migration with regard to destination choice as well as adaptation and acclimatisation at the place of destination. The interviewees relied greatly on social ties with their family and co-ethnic networks in business start-up and development, including receiving business advice, having members of their family and co-ethnic networks as staff, and enlisting assistance from spouses and relatives living in cities during highpressure working hours (e.g. Table 2).
To address the second research question, a novel aspect for studies of rural immigrant entrepreneurs with refugee backgrounds is their ability to maintain material and emotional relationships with their countries of origin over their life course. We found that business decisions and migrant trajectories were closely related to transitions in the restaurateur's family, for instance when their children started to receive higher education or when the restaurateur approached retirement. Many interviewees had started their restaurant to offer their family and children better housing and opportunities for education. They were considering pausing or stopping their work as restaurateurs when their children needed support for their leisure activities or when they started to study. Some even stated that the whole family would follow the children if they moved to study in larger urban areas such as Stockholm, yet they did not rule out a return to their current rural place. Most interviewees' social networks included ties with people who owned an apartment or a house in or near their origin country, which the interviewees could use for business or pleasure. Towards retirement, some of them said they would like to live in Sweden during the summer and in their origin country or in a warm place during the winter. Hence, the interviewees' transnational social networks and multiplex geographic, business, and institutional embeddedness enabled their flexible attitudes towards mobility based on a plethora of possible places for living, working, holidays, or seasonal migration. This extends our understanding of the fluidity of globalising rural places through fleeing and returning, and how such complex migrant trajectories may have an impact on social networks (Mitchell 1969;Milbourne & Kitchen 2014).
In sum, we have advocated a relational approach to globalising countrysides following our study of relations between immigrant restaurateurs with refugee backgrounds and their fluctuating proximate and distant family and co-ethnic networks. Through these individuals' and groups' social and economic relations, we have examined how the global and the local reshape each other, and how this may contribute to unequal geographies of globalisation. The relational approach used here has illuminated how individuals influence, and are influenced by, conditions in multiple places through local, regional, and transnational place-to-place mobilities on a daily basis and over the life course. This offers novel insights into spatial relations (i.e. regional and transnational urban-rural links) to existing studies of migrant entrepreneurship that may stimulate future research on shifting power relations and social constructions in globalising countrysides.