Poverty, networks, resistance: The economic sociology of Roma migration for begging

Migration for begging and informal street work in Western Europe has become a common livelihood strategy in many Romanian Roma communities, and over the last decade, Scandinavia has emerged as an important new destination. Using a combination of quantitative survey data on Romanian migrant beggars and street workers in Oslo, Stockholm and Copenhagen, and qualitative ﬁeldworks in Scandinavia and Romania, this article presents a framework for studying migration for begging that goes beyond widespread narratives rooted in either culture or poverty. I argue that migration for begging is a purposeful economic adaptation, embedded in three distinct sets of social phenomena. First, the social and economic processes of marginalization of Roma communities in post-socialist Romania can help explain the motive for migration in terms of poverty and lack of alternative options. Second, the structure of social capital within Roma households and communities can help explain why they are able to engage in transnational migration under extremely difﬁcult conditions despite lacking economic and educational resources. Third, ‘oppositional’ Roma identities can help to explain why some are willing to engage in ‘transgressional’ activities that others perceive as shameful, thus allowing the exploitation of marginal economic resources in times of economic hardship.


Introduction
The Roma remain Europe's most disadvantaged and ostracized minority. With an estimated 1,850,000 people, or about 8.6 percent of its population, Romania has Europe's largest Roma population and since the collapse of communism their social situation has remained largely unimproved or even deteriorating. With Romania's accession to the EU and the opening of borders to the West, many have turned to migration for survival. However, unlike the millions of other Central and Eastern Europeans who have moved west over the last decade in search of employment, many Roma pursue income through begging and informal 'street work'. As EU citizens their presence is not formally challenged, but since they are outside the formal labour market they are effectively without access to social rights and pathways to permanent residence (Tervonen and Enache 2017). As a result, the plight of the Roma has become highly visible on the streets of Western European cities. Perhaps nowhere is the contrast to the affluence of their hosts more striking than in the rich Scandinavian welfare states. Since 2007 a steady stream of Roma migrants have travelled to Scandinavia using regular bus services, private cars or informal shuttle buses linking migration hot-spots in the Romanian countryside with Scandinavian cities, in a transnational field where individuals and families travel back and forth on a frequent and regular basis. Sleeping in parks, on pavements, in parked cars or camped in the forest, they beg, collect bottles, play music or perform minor services, recycle scrap metal or engage in petty crime. Relative to other migration flows their numbers are modest. Estimates vary and populations fluctuate according to season, but the total numbers of migrant street workers have hardly exceeded a few thousand people at any time in any of the three countries. Nevertheless, their conspicuous presence in public spaces has created heated political debates. As elsewhere in Europe opinions over how best to respond are polarized between calls for security and control and appeals to fight discrimination and promote social inclusion. Despite overall institutional similarities the Scandinavian countries have opted for noticeably different policies. Denmark has adopted the most rigorous approach to deter their arrival, with a national ban on begging selectively applied to foreigners, harsh police tactics against rough sleepers and strict limitations on public NGO funding for services to homeless migrants. Sweden has taken the opposite position, allowing migrants to beg and sleep in public spaces and allocating NGO funding for basic services, shelter and food. Norway has taken an intermediate position where the city of Oslo imposed a municipal ban on sleeping outdoors; a national ban on begging was proposed in 2014 but withdrawn the following year; and NGOs receive ad hoc funding for services. Nevertheless, all three countries attract significant numbers of migrant street workers (Djuve et al. 2015). There is today a growing body of research on the racialization of Roma migrants in public discourse (Yuval-Davis, Wemyss and Cassidy 2017), how Roma migration is framed as a threat to public order and security and the subsequent rise in anti-gypsy sentiments throughout Europe (Pusca 2010;Nacu 2012;Sigona and Vermeersch 2012;Vermeersch 2012). Less scholarly attention has been devoted to the underlying forces that shape these migration flows. Migration for begging-as a form of economic mobility outside formal labour markets yet within the framework of European free movement-constitute a distinct form of migration in today's Europe. While some argue that both 'nomadism' and begging are practices rooted in Roma culture, others maintain it is a desperate response to poverty and discrimination. I argue, however, that such a stark division between economy and culture is unfruitful. The purpose of this article is instead to develop an empirically grounded analytical framework for understanding Roma marginal migration and begging, rooted in economic sociology. The overall research questions are: Why has begging in Scandinavia become a common livelihood strategy in many marginalized communities in Romania-and why is it almost exclusively practised by the Roma?

Migration for begging as embedded practice
In public discourse, as well as policy documents and legal systems across Europe, both mobility and begging is often described in terms of 'Roma culture'. In 2008, for example, the Italian Supreme Court stated as a mitigating circumstance, in a case involving child begging, that begging is a 'Roma cultural practice', and 'a traditional way of life deeply rooted in the culture'. 1 In Norway, Article 5 in The Council of Europe's Convention for the Protection of National Minorities regarding the right 'to maintain and develop their culture', has been interpreted to the effect that Norwegian Roma should be able to maintain a semi-nomadic lifestyle (Tyldum and Friberg 2014;Engebrigtsen, 2015). The notion that both nomadism and begging are practices rooted in the traditional culture of Roma can find some support in the academic literature. While the majority of Eastern Europes' Roma today are sedentary, scholars often describe the traditional adaptations of the Roma in terms such as 'commercial nomads' or 'service nomads', specializing in the mobile supply of various trades and services to the surrounding populations (Hayden 1979;Nemeth 1986;Gmelch and Gmelch 1987). In his introduction to Roma culture, Yaron Matras (2011) writes that the Roma 'prefer not to work for others as employees, as far as possible' and that they 'tend not to recognize any fundamental division between productive-manufacturing skills (. . .), and strategies that are regarded by outsiders as "parasitic," such as begging' (Matras 2011). Fraser (1995 notes that some of the earliest written sources about the Roma from the fifteenth century, mention begging practised by Gypsies presenting themselves as Egyptian pilgrims (Fraser 1995). Roma begging has even been described as the activity of the last 'hunter-gatherers' in the West, whose objects of gathering no longer consists of natural products of the land but the cultural products of surrounding societies (Formoso 1986).
Roma activists and NGOs have been far more reluctant to invoke notions of culture. In the case of the Italian Supreme Court ruling, for example, Roma activists argued that begging does not reflect traditional Roma culture, but marks its collapse and crisis. According to them, begging is a more recent phenomenon, following the economic transformations which led to the disappearance of traditional occupations previously held by the Roma, such as horse raising, circuses, iron smithing, etc. and the subsequent marginalization of many Roma communities (Ruggiu 2016). There is a vast body of research carried out by NGOs and European agencies which supports the argument that today's pattern of migration is a result of age-old discrimination and exclusion (FRA 2009;Van Baar 2011. According to a report from a large-scale research project commissioned by the UNDP, for example, migration is an 'option of last resort' for the Roma. They conclude that the reasons why Roma migrate do not differ from those of non-Roma, namely poverty, lack of jobs and desire for a better life. What distinguishes them from other mobile EUcitizens is the extreme poverty that drives them and the hostile responses they provoke in their destinations (Cherkezova and Tomova 2013).
In this article, I argue that this distinction between culture and poverty is unfruitful. Consider, for example, one of the few in-depth ethnographic studies on the subject. Catalina Tesar (2015) who followed a group of Romanian Cortari Gypsies in Italy, describes begging as a form of performative work which requires bodily training and specific skills.
She describes how the Cortari, when crossing the border, ritually change their customary dress for worn out clothing, along with their bodily posture and hygiene. The purpose, she argues, is to draw strict symbolic boundaries between home-structured by kinship, gender and morality-and life abroad-defined by economic accumulation through begging. However, does the need to strictly separate between the 'real' life at home and the instrumental economic activities abroad suggest that begging is not a part of Roma culture? Or, is the ability to morally separate so strictly between the two a cultural precondition for begging?
In the following, I propose to study migration for begging among Romanian Roma within a framework of economic sociology using the concept of embeddedness. First developed by Mark Granovetter, the concept of embeddedness emphasizes that although individuals' actions may be both rational and goal-oriented, the goals people pursue, and the means by which they pursue them, are shaped by the social relationships, institutions and structures in which they are embedded (Granovetter 1985). This means that economic action are both constrained and enabled by reciprocity expectations built up through interaction, and that the pursuit of material gains may interact with other goals such as a quest for approval and social status (Portes 2010). Moreover, economic practices tend to become institutionalized, e.g. infused with value and meaning beyond the technical requirements of the task at hand. Institutions, which may be defined as symbolic blueprints guiding relationships between roles (Giddens 1993;Portes 2010), usually arise as adaptive responses to environmental, social and political pressures. But once infused with value and meaning, they tend to become 'sticky', continuing to shape behaviour in changing environments, irrespective of their original purpose (Fukuyama 2011: 452). Finally, economic transactions tend to be shaped by unequal power relationships between groups and individuals (Portes 2010). In this perspective, it makes little sense to make strict distinctions between the realms of economic behavior and that of culture.
How can this perspective be applied to the study of migration and begging among today's marginalized Roma? The anthropological literature on Roma economic adaptations often underscore the wide diversity in adaptations between the various groups labelled as Roma and begging is no exception. Far from all Roma groups engage in begging and among those who do, the frequency, function, and meanings attributed to it differs widely. There are, however, certain basic social and institutionalized patterns which are based on historical experiences of slavery, persecution and economic marginality, that appear to be common across different groups of Roma and which may be relevant for understanding both migration and begging.
The first such pattern is a deeply felt and ritually confirmed separation between the world of the Roma and that of the non-Roma (Gadjo), based on practices that mark the strict differentiation between the two (Formoso 1986;Horváth 2005;Stewart 2013). The particular content of these practices vary, but typically include a complex patriarchal ideology of gender differences and elaborate rituals and taboos regarding purity and pollution (Sutherland 1986;Blasco 1999;Troc 2005;Engebrigtsen 2007). The second is the development of a mode of subsistence outside of state and wage labour, carved out within asymmetric fields of socio-economic relations intertwined with, as well as dependent on, dominant societies (Stewart 2013;Brazzabeni, Cunha, and Fotta 2015). Despite huge variations in the economic adaptations of various Roma groups, most share a preference for autonomy and self-employment, often rooted in conceptions of what is considered as real 'Gypsy work' (Stewart 1997;Formoso and Jean 2000). The third is a deep-seated sense of belonging to family and kin and a sense of personal and collective autonomy and superiority as Roma (see, e.g. Okely 1983;Formoso 1986;Stewart 1997;Horváth 2005;Engebrigtsen 2007). In an attempt to preserve not just continuity and distinction, but also a sense of self-worth in a situation of geographical and economic immersion into hostile and more powerful majorities, the Roma continually try to make themselves appear superior by subverting the surrounding social and moral hierarchies through ritual taboos regarding cleanliness or notions of real gypsy work (Stewart 2013). These distinctions are not absolute and several studies underscore how many Roma groups rather than general resistance and innovation tend to incorporate some elements of the surrounding culture while resisting others. For example, Jan Grill describes how a group of migrant Roma from Slovakia in the UK oscillate between paid physical labour and various ways of 'fixing up money' through skillful manoeuvring within different state systems (Grill 2012(Grill , 2015.
Regarding migration and begging, however, two implications may be drawn from this. The first is that although nomadism is long abandoned by most Romanian Roma, the exploitation of marginal resources outside formal institutions-whether as traders, itinerant craftsmen, entertainers or seasonal labourers-necessarily involves a significant degree of mobility, relying on networks of family and kin rather than formal institutions. With open borders and low-cost transportation, transnational migration may thus not be very different from going to a neighbouring country in search of casual work, scrap-metal for recycling or a good trade deal. The second has to do with the moral economy of begging and street work: begging may not have any cultural significance as such, but by maintaining a relatively self-contained value system and strict boundaries between insiders and outsiders, many Roma obtain a certain degree of indifference to the legal, moral and aesthetic concerns of the majority. This, in turn, can provide at least some measure of psychological protection against shame and stigma inflicted by outsiders.
In the following a basic empirical framework for understanding migration and begging among Romanian Roma as an economic adaptation is outlined. This framework consists of three elements; motives-in terms of the poverty and marginalization; resources-in terms of social capital based on household and kinship; and skills-in terms of the ability to endure the stigma associated with begging.
Motive: Sleeping in make-shift forest camps, dodging police and security guards, begging for money and rummaging through trash bins looking for bottles in near-arctic conditions, is extremely demanding, it involves considerable risk, and it generates rather modest incomes. The first step in our analysis is therefore to explore the specific political and economic process of marginalization that have left the Roma with so few alternatives-both in Romania and in alternative destinations in Southern Europe-as to make migration for begging in Scandinavia appear an attractive option.
Resources: Although poverty and marginalization may be powerful motives, it is not a very good predictor of migration, because migration is risky and takes significant investment and resources (Czaika and de Haas 2012;Carling and Talleraas 2016). In Bourdieusian terms it depends on access to economic, cultural and/or social capital (Bourdieu 1986). Most rural Romanian Roma have little savings and limited access to formal credit; nor do they have much education or formal skills; and they are often excluded from networks that could link them to the formal economy. The second step in our analysis is therefore to explore how the practice of marginal migration for begging and street work is embedded in a particular form of social capital based on support networks of bounded solidarity and enforceable trust within families, villages and communities, which can be mobilized as resources in the pursuit of marginal economic opportunities.
Skills: Motive and resources, however, are not enough. For most poverty-stricken Romanians, going to Scandinavia and begging in the streets is almost inconceivable; not because of the physical hardship and risk involved, or the relatively meagre outcomes it produces, but because of the moral stigma and shame it would inflict. The third step in the analysis is therefore to explore how oppositional identities based on a cultural resistance against mainstream society's continuing attempts to define the Roma as inferior, serves to protect against such stigma. The concept of 'oppositional identities' was originally developed to explain academic disengagement among African-American youth, who, as a response to racial discrimination, were prone to see adherence to white middle-class norms as a betrayal of black culture and instead would place value on alternative norms based in 'street culture' (Ogbu 1995;Ogbu and Simons 1998). Willis (1977) described the same mechanism among white British working class 'lads' who developed a culture of resistance to academia and authority, protecting their sense of dignity and self-respect in the face of repressive class structures. Portes and Rumbaut have developed a parallel concept of 'reactive ethnicity' to describe how children of disadvantaged immigrants developed discourses and self-images negative towards mainstream institutions (Portes and Rumbaut 2001). A similar culture of resistance lies at the heart of Roma identity. As a way of dealing with their position at the bottom of the social order, adapted to centuries of slavery, persecution and marginalization, the Roma have developed customs, practice and identities that are defined directly in opposition to the mainstream societies surrounding them (Stewart 1997). This has a double function: to impose and maintain strict boundaries between insiders and outsiders, and to make possible an inner cultural space where the world of the Gypsy and their life-styles are perceived as more desirable. This moral boundary provides some psychological protection against stigma, which in turn enables people to exploit marginal opportunities that others in similar circumstances are more likely to shun. In combination, I argue, these three elements go a long way to explaining why begging in Scandinavia has become a common livelihood strategy in many marginalized communities in Romania, as well as why begging is almost exclusively practised by the Roma.

Data and methods
The analyses combine quantitative and qualitative data from a large research project on Romanian migrant street workers in Scandinavia (Djuve et al. 2015). The quantitative part consists of three separate surveys of 1,269 migrants in Stockholm (n = 446), Oslo (n = 438) and Copenhagen (n = 385), conducted during the summer and autumn of 2014. We used Respondent Driven Sampling (RDS), a method utilizing social networks to produce data and representative estimates for hard-to-reach populations. Originally developed for public health and HIV research in high-risk populations, RDS has in recent years been extended into migration studies (Tyldum and Johnston 2014). It builds on snowball sampling, but POVERTY, NETWORKS, RESISTANCE 233 incorporates methodological and statistical elements to mitigate biases (Heckathorn 1997;Tyldum and Johnston 2014). RDS sampling is based on a double incentive structure whereby respondents-who remain fully anonymous-are paid both to take part in interviews and to recruit new respondents. The combination of economic incentives, peer recruitment and anonymous participation provide access to population groups and information that would otherwise be difficult to reach. 2 To become eligible for participation, respondents had to be from Romania and not have a regular job or place to live in Scandinavia. The sample thus consists of both Roma and non-Roma migrants who were homeless and formally unemployed while in Scandinavia. Interviews were conducted in locations specifically adapted to organize interviews and great care was put into creating an environment conducive to establish trust among participants. In Oslo and Stockholm we rented space from local churches, while in Copenhagen interviews were conducted in a rented store-front location. The interviews consisted of an almost hour-long questionnaire, verbally administered face-to-face by a team of Romanian and/or Romani speaking interviewers.
The qualitative data was collected during three separate fieldworks in Romanian sending regions (Gorj and Buzau Counties) and several shorter field-works in the three Scandinavian capital cities. A substantial amount of qualitative data was also collected in conjunction with the quantitative surveys, which required close monitoring and contact with the target population. Roma and non-Roma migrants were the primary targets for all qualitative interviews, but also family members, non-migrant villagers, social workers, teachers, municipal leaders, police and NGO workers were interviewed, both in Romania and Scandinavia. Some interviews were recorded and transcribed, and lasted for hours, while others were shorter informal conversations. More than 180 people were interviewed in the qualitative part.

Measuring ethnic identity
As already noted, Roma populations differ widely in their adaptations and social situation (Stewart 2013). This also includes the extent to which they have maintained a distinct Roma identity. While some maintain a traditional and separate Roma lifestyle, others are largely assimilated into the surrounding populations. In common language used by Roma and non-Roma alike, people are referred to as being more or less 'traditional' or 'romanianized'. 'Traditional' Roma usually refers to people who live segregated from the majority and maintain the use of Romani language as well as traditional cultural practices regarding dress, food and family relations. 'Romanianized Roma' usually refers to people who have abandoned the use of the Romani language and traditional clothing, but who have not attained full acceptance by mainstream society. The degree of assimilation or traditionalism may vary considerably both at the community and individual level and, in the quantitative analysis, this variation will be used in order to identify to what extent various social phenomena are related to the strength of Roma identity. We use self-identification and language as indicators of ethnic identity: Self-identity: 41 per cent of survey respondents refer to themselves as Roma or Gypsy; 16 per cent prefer the term 'romanianized' Roma; 35 per cent call themselves ethnic Romanians; and another 8 per cent prefer other labels (mostly Turkish or Hungarian). Language: 51 per cent report that they speak Romani fluently; 6 per cent report some knowledge; while 43 per cent do not speak Romani at all. 33 per cent use Romani at home with their families; 14 per cent uses a combination of Romanian and Romani; 48 per cent use Romanian only; while another 5 per cent uses some other language (Turkish or Hungarian). Based on a combination of self-identity and language, we may construct three categories within our sample: Those who identify as ethnic Romanians and who report that they do not speak or understand Romani language are categorized as 'majority Romanians' (n = 319). Those who identify as Roma or gypsy and who report that they know Romani language and uses it at home with their families are categorized as 'traditional Roma' (n = 516). The residual then consists of three categories: (1) People who prefer to call themselves 'romanianized Roma', and who may or may not speak Romani language.
(2) People who identify as ethnic Romanians, but who report that they know and/or use Romani language. (3) People who identify as Roma or Gypsy, but who do not know Romani and/or use Romani language. Respondents belonging to any of these categories will be labeles as 'romanianized' or assimilated Roma (n = 432). In the following, the differences between 'majority Romanians', 'romanianized' or assimilated Roma and 'traditional' Roma, will be used to show how various social phenomena are related to ethnic identity.
We start with how many in our sample who have engaged in various income seeking strategies during the last week before the interview. Table 1 show only minor differences between the three categories when it comes to how many who collect bottles (a large majority in all groups), sell magazines (a significant minority in all groups) and who play music or sell small items in the street (a relatively small minority in all groups). There are, however, two significant differences between groups: First, non-Roma are much more successful in obtaining casual work. Second, traditional Roma are much more likely to engage in begging. While only 18 per cent of those labelled as non-Roma report having begged for money during the last week, the share increases 57 per cent among the assimilated Roma and 77 per cent among the traditional Roma.
In the following, I will present three explanations for why migration for begging has become a livelihood strategy followed by Roma in particular.

Poverty and marginalization: the Roma as post-socialist 'transition losers'
The first step in order to understand migration and begging as a livelihood strategy is to explore the increasing marginalization of Roma communities in Romania, which over the last few decades have left them with few other alternative options both at home in Romania, POVERTY,NETWORKS,RESISTANCE 235 as well as in the formal labour markets of their potential destination countries. We can identify three such distinct yet interrelated processes.
First, many of the Roma's traditional economic niches have dwindled in importance or even entirely disappeared. Throughout most of their history the Roma have subsisted on specialized services and occupational niches and Roma identity was associated with hereditary artisanal crafts, such as brick-making, basket-weaving, tin-and coppersmiths, horse-trading, bear-taming, etc. Over the last century, however, this craft-knowledge has slowly disappeared, partly as a result of state assimilationist policies during socialism, partly as a result of a modernizing economy. During socialism, the Roma were mobilized to work in state agriculture and manufacturing, but after the revolution, they were the first to lose their jobs. The Roma have also served as an informal manual labour reserve in agriculture, industry and public works, but with the mechanization of agriculture and restructuring of industry and manufacturing during post-socialism, the need for manual labour has fallen sharply. Second, following economic restructuring, ethnic boundaries became increasingly significant. Under communist rule, ethnic identities were seen as relics of the past and obstacles to achieving the new society. Traditional markers of Roma identity were repressed, but so was overt discrimination. After the revolution, the Romanian majority sought to redefine its national identity -often in contrast to the Roma 'other'. Surveys throughout Central and Eastern Europe, show that hostility towards Roma increased significantly in the post-socialist era (Csepeli, Ö rkény, and Székelyi 2000;Barany 2002). As ethnic boundaries between Roma and non-Roma became increasingly salient, discrimination and exclusion was intensified (Stewart 2012). Third, many Roma hold a deep-rooted scepticism towards formal education. Combined with discrimination in schools, the result is that education levels remain low, and many children hardly attend school for more than a couple of years. Adjusting to new economic realities by entering the formal economy is therefore extremely difficult.
The recent economic history of the Roma in Gorj County, a major sending area for migrants to Scandinavia, can serve as an illustration. As elsewhere in Romania, older Roma in Gorj tend to remember communism as a golden era, when they were provided income and employment in collectivized agriculture and state-owned mining and manufacturing. To them, the revolution marked the beginning of a gradual downward spiral. The mining industry was shut down in the early 1990s. At the same time, state-owned lands were privatized and distributed to those who could document a historical claim as peasants. The Roma-having been slaves until the mid-19th century and excluded from holding land-were thus excluded from the privatization process. Many continued to work as agricultural labourers, but mechanization and restructuring rapidly diminished the need for manual work. Most Roma in Gorj County are so-called Caramizari-traditional brickmakers-and during the 1990s they turned to their traditional craft, selling bricks for the local market. After the millennium, however, new factory-made bricks were introduced, driving traditional brick-makers out of business. Casual work, or 'Zillier'-which typically involves well-connected individuals setting up work-teams who travel the country to dig ditches, gravel roads, etc.-was, and still is, another common source of income. But like in agriculture, new technologies have drastically reduced labour demand. In the early 2000s many started migrating to Italy, where some would beg while others found work in agriculture. With the ongoing economic crisis and increasing competition over informal employment from African immigrants, opportunities for work in southern Europe largely disappeared. When Romania joined the EU in 2007, few alternatives remained. Pioneers sought new destinations to the north, and within few years, going to Scandinavia for begging and street work was the primary income-source in many villages. Regarding children's schooling, most parents interviewed recognized the benefits of acquiring basic literacy and English language skills, but many are ambivalent, and few have any hopes that an education will help access employment or income. Discrimination against Roma children is prevalent: most villages have a main school building in the centre, and smaller separate school premises near the Roma settlement, where Roma children receive separate schooling under inferior conditions. While school principals justify these arrangements as a means of lowering the threshold for attendance, such arrangements are in reality initiated by non-Roma parents and teachers in order to avoid unwanted contact. Roma villages and settlements throughout Gorj County, although often lacking in basic public amenities, is today dotted with newly refurbished houses paid for with money from begging in Scandinavia. Given their present situation, the Roma of Gorj appear to have few alternative options.
Our quantitative sample consists of migrants on the streets in Scandinavia without regular employment or a regular place to sleep, and most respondents are poor, even by Romanian standards. Nevertheless, using measures of literacy, basic education, employment history and housing quality as indicators, there is a striking relationship between ethnic identity and marginalization (see Fig. 1). Among the non-Roma, 89 percent report that they can read and write in Romanian; 85 percent have completed 8 years of mandatory education; 65 percent have had formal employment in Romania; and 62 percent report that their house in Romania has piped water inside. Among the traditional Roma, only 28 percent can read and write in Romanian, only 17 percent have completed 8 years of schooling, only 14 percent have ever had formal employment (mostly elders who worked in state mining etc. during communism), and only 22 percent have piped water inside their house in Romania. The so-called assimilated Roma lies somewhere in between on all measures. In light of this, it is not unreasonable to see migration for begging to some extent as a desperate response to a situation of extreme poverty and deprivation.

Family structure and village networks as migration capital
Although poverty may be a powerful motive for migration, this is not enough, since actual migration takes resources and investment. Poor people are often cut off from the possibility of migration, but this does not seem to apply to the Roma to the same extent. The second step in the analysis is therefore to explore what alternative resources they possess, which can be mobilized for migration purposes. I will show that these resources are best conceptualized as social capital based on family structure and village networks based on common identity. Consider the following excerpt from field notes collected at an emergency shelter in Oslo: A large forest camp outside Oslo was dismantled by police in the middle of winter. The police confiscated tents and sleeping bags, leaving more than 60 migrants with nowhere to sleep in sub-zero temperatures. The NGO-run emergency shelter was overwhelmed, and people had to draw lots. A green token meant you could enter; red tokens meant you were refused. Most people drawing red tokens would complain, but quickly move on in small groups. However, one elderly woman appeared to be in particular distress, sitting motionless on the pavement and crying before moving off. Others in the line appeared to pity her. A Roma woman explained: 'We will always manage; we stick together and look out for each other. For her it is different. She is a Romanian, she doesn't have anyone' The fact that informal networks are an important source of social capital for migrants is one of the most widely acknowledged insights from the sociology of immigration, and by no means exclusive to the Roma (Massey et al. 1999). However, as the note above suggests, the particular structure of social capital within Roma families and communities differs from their non-Roma neighbours in ways that affect how they can be mobilized as migration-related resources. This can be conceptualized at two levels. First, Roma tend to have larger households, due to early marriages and generational housing, allowing spouses to travel together, while other household members take care of the children (Pantea 2012). Many respondents emphasize how travelling as a family unit provides invaluable companionship and support in the face of hardship and risk. Second, dense identity-based village networks provide a strong source of solidarity and mutual support within a wider community of migrants, which is effectively used to access transportation, protection, and places to sleep, companionship and emotional support, economic opportunities etc.
The importance of family and village based social capital may be illustrated by how it regulates access to begging spots. Having a good spot-e.g. outside a busy subway or shopping centre-can mean the difference between failure and success and is essential for maintaining a continued pattern of migration. However, good spots are scarce and in order to make a profit one must avoid overcrowding. Different areas of the city tend to be dominated by people from the same villages and these networks are an important source of support. Individual spots within these areas tend to be controlled by individuals and families. Sometimes family members rotate so that when one person goes back to Romania another family member uses the spot and thus makes sure no one else takes it. In places where competition over spots was particularly intense, people told us about extortion and attempts to monetize spots. However, most places were large village networks operated within the same areas so people tried to avoid conflicts over the use of spots and disputes could sometimes be resolved by the village tribunal back home. It should be noted that while most migrants saw tight-knit families, kinship groups and village networks as an invaluable resource for protection and support, dependency on such networks also render migrants vulnerable to abuse. Many borrow money for the journey and control of resources that people depended onsuch as a car, a place to sleep or a spot to beg-can lay the basis for highly exploitative relationships, within or outside the family. But as Pantea (2013) points out, this is usually a greater risk for those who are disconnected from trustworthy and resourceful migration networks.
The particular structure of social capital in the form of dense networks of solidarity and support can be illustrated using quantitative data on household size, group size with who one travels and group composition, e.g. whether one travels with close family members or not (Fig. 2). The figure shows that compared with the non-Roma, the Roma-and in particular the more traditional ones, tend to have larger households; they tend to travel in larger groups; and they are far more likely to travel with close family members. When these networks and relationships are mobilized for migration purposes, I argue, they can to some extent substitute economic and formal resources which the Roma often lack.

'Oppositional identity' as protection against stigma
So far we have discussed resources in terms of the structural poverty that motivate migration and the social capital which enable it. However, economic migrants also need to generate income. Begging, the primary strategy employed by Romanian Roma in Scandinavia, not only requires bodily training and specific skills (Tesar 2015); as Thomassen points out it is also a form of exchange in which, due to its marginal and liminal nature, central norms regarding morality and worth are being negotiated (Thomassen 2015). While gift-giving invokes a logic of reciprocity and serves to reinforce social bonds in everyday life, begging is by nature non-reciprocal, in the sense that the giver cannot expect any return, except a display of moral gratitude and perhaps a 'feel good experience' (Mauss 1966). Such alms with no expectation of return can become 'poisonous' leaving the recipients stigmatized and in 'moral debt' (Parry 1986). This stigma, I argue, represents a major barrier against begging as an economic strategy. Consider, for example, the following excerpt from field notes in a migrant hotspot outside Buzau in Romania: Speaking to a local non-Roma woman-highly knowledgeable about the intricacies of travel and finding spots to beg and places to sleep in Stockholm-it became clear that she was quite envious of her migrant neighbours. 'I would go tomorrow, I really would!' she said repeatedly. When I asked her why she did not go, she simply replied that 'no, no, it is impossible; it is not for me'. As I kept pushing the subject, and she kept dodging the question, my assistant finally cut me off: 'Come on, stop pressing her, you know very well why she can't go! She is not a Gypsy'.
The reason why this woman could not go was not a lack of resources or of networks, but a question of morality and identity; she was simply not 'the kind of person who begs' and if she did it would have serious repercussions for her social standing and sense of self-worth. Family and network resources, by ethnic identity: Share who report that their household in Romania consists of more than six people; that they travelled to Scandinavia in a group of four people or more; and that they are staying in Scandinavia together with close family members. Marginal migrants from Romania in Oslo, Stockholm and Copenhagen. N = 1,266.
Because for non-Roma villagers living in migration hot spots, the most typical reaction when discussing migration and begging was not envy, but outrage over what they saw as immoral behaviour: We know what they are doing there! They are begging for money, stealing, and doing all kinds of things. They have no shame, the way they act. It is a shame for the whole village.
Migrant beggars provoke considerable hostility in Scandinavia too, something of which they were well aware. In the quantitative survey a majority of beggars reported experiences of being chased from public spaces, refused access to shops, being shouted at and spat on in the streets, etc.
The Roma are by no means immune to the denigration and contempt they face both at home and in Scandinavia. But the Roma identity and sense of self-defined in opposition to the non-Roma-represents a defiant resistance against incorporating the majority's view of them into their own. Stewart (1997) paraphrases Günter Grass' notion of 'gaiety in the face of despair', to explain the 'core' of Roma culture, as well as the fury they often provoke in others. Such 'oppositional identities' are perhaps better explored through in-depth ethnographies than more formal qualitative interviews. Yet the 'cultivated insouciance' and 'careful disregard and attentive disdain for the non-Gypsy way' (Stewart 1997: 12) tends to shine through, as in the way one young woman shrugged off my questions regarding the opinions of their non-Roma neighbors: Yes, of course they look down upon us [for going abroad to beg]. But they look down upon us anyway, no matter what we do. What do they know? They don't know what it is like to be poor.
Many Roma seemed to have few illusions or ambitions of ever gaining the respect of the 'Gadjo', and many would simply dismiss their opinions as expressions of envy. 'They are just envious because they don't have the same spirit as the gypsies', as one elderly man put it. A few told us how they felt shameful for begging, and that they preferred more respectable means of income, such as collecting bottles or selling magazines. But most informants would talk about begging in a detached and instrumental manner. Non-Roma street workers would often go into lengthy discussions of shame and self-worth, and elaborate attempts to present themselves as respectable persons despite their situation ('I do not want to live like this, I am a good worker', etc). Roma street workers would more often go into detailed comparisons of how much one could earn per day, and what were the best spots or time of day to beg. And while most beggars and street workers had experienced instances of harassment and hostility from locals in Scandinavia, the Roma often downplayed their significance, insisting that it was no big deal and that the hostility was probably justified. Most Roma beggars seemed to have just as few ambitions or illusions of gaining the respect and recognition of people in Scandinavia. This stood in stark contrast to majority Romanians, who were often deeply concerned about the moral standing of Romanians in the eyes of Scandinavians, never missing an opportunity to underscore that the gypsy beggars are not 'real Romanians'.
In many ways, oppositional identities resemble the fox and the sour grapes: Unable to reach the grapes that hang above him, the fox declared them sour, because after all, it is better not to want the things you cannot get (Elster 1983). The grapes in this case being POVERTY, NETWORKS, RESISTANCE 241 moral recognition from the surrounding majority. This moral separation between the world of the Roma and the world outside and their resistance against incorporating the majority's view onto themselves provides at least some protection against the 'symbolic violence' of mainstream society's contempt. In turn, it allows them to pursue economic strategies others perceive as too shameful.
Measuring 'oppositional identities' in surveys is difficult, but we may use two different indicators as proxies. First, we measure adherence to the traditional prohibition within Roma society against women wearing trousers in public. As Fig. 3 shows, only 6 per cent of the non-Roma reports that they find it unacceptable for women in their family to wear trousers. Among the assimilated Roma, the number is 33 per cent, while as much as 72 per cent of the traditional Roma find this unacceptable. As one of the more tangible ritual taboos which mark the boundary between Roma and non-Roma, I argue that adherence to this taboo is a good proxy for Roma oppositional identity. Second, we may use a direct measure of attitudes towards begging. Respondents were asked to indicate if they agreed, disagreed or did not have an opinion regarding the statements: 'Begging is just as good as having a job-as long as you put food on the table' and 'It is humiliating for a decent person to beg.' Figure 3 shows the shares that indicate an opposition to mainstream norms regarding income earning by either agreeing to the first statement, or disagreeing with the second. Once again we find an ethnic dimension, from 29 per cent among the non-Roma to 61 per cent among the traditional Roma.

Can poverty, networks and resistance explain why the Roma beg?
So far, the quantitative data has showed that structural marginalization, family social capital and oppositional identities all appear to display a distinct ethnic dimension, in the sense that Figure 3. Indicators of 'oppositional identity', by ethnic identity: Share who report that they think it is unacceptable for women to wear trousers in public; and share who agree to the statement 'begging is just as good as having a regular job, as long as it brings food to the table' and/or disagrees to the statement 'it is humiliating for a decent person to beg for money'. Marginal migrants from Romania in Oslo, Stockholm and Copenhagen. N = 1,266. they are more commonly found among Roma than non-Roma, and that they are more commonly found among the more traditional Roma than among the more assimilated Roma. The question, however, is to what extent these concepts can explain the practice of transnational begging. Since we do not have quantitative data on the population of potential migrants in sending regions in Romania, we cannot say anything about why some people migrate while others stay home. However, we may say something about why some people pursue begging as an income strategy once abroad, while others follow alternative economic strategies. A multivariate logistic regression using a dummy variable on whether or not having begged for money during the last week as the dependent variable may provide an answer. Table 2 shows the results. In the first step of the model, we introduce the control variables, which include gender and city, as well as Roma identity as predictor variables. It shows that within the population of Romanian migrants living on the streets in Scandinavia: (1) women Step 1 Step 2  are far more likely to beg than men.
(2) Begging is most commonly pursued in Stockholm (where policies towards beggars are most lenient) and least common in Copenhagen (where begging is illegal).
(3) Most importantly for our purposes: Roma are far more likely to beg than non-Roma, and the more traditional Roma are far more likely to beg than the more assimilated Roma. In step 2, five new explanatory variables are introduced. As indicators of structural deprivation, I use a dummy variable indicating illiteracy (cannot read or write) and whether or not the respondents' house in Romania has piped water inside. As an indicator of family based social capital, we use a dummy variable on whether or not the respondents' household in Romania consists of more than six people. As an indicator of 'oppositional identity', we use a dummy variable indicating whether or not the respondent find it unacceptable for women in their family to wear trousers in public, and whether they agree that 'begging is just as good as job' and/or disagree that 'begging is humiliating'.
The results in step 2 shows that both indicators of poverty have a significant effect on likelihood of begging: those who cannot read or write are more likely to beg than those who can, and those who lack piped water inside their house in Romania. Family social capital also has an effect, as those who have large households are significantly more likely to beg than those belonging to smaller households. Both indicators of oppositional identity also have a significant effect on begging: those who report 'oppositional' attitudes towards employment and begging are not surprisingly more likely to beg than those who report more mainstream norms. This, of course, may be a case of adaptive preferences, as people who beg for other reasons (such as extreme poverty) may try to cope with their situation by adapting their attitudes. This is analogous to how the concept of oppositional identities itself refers to a form of adaptive attitude. However, even when controlling for attitudes towards begging, we find that those who find it unacceptable for women to wear trousers are more likely to beg than those who do not, suggesting that norms matter beyond mere attitudes towards begging as such.
An important result is that once indicators of structural poverty, family social capital and oppositional identities are introduced in step two, the effects of the original variables changes. The effect of gender is only slightly reduced, which reflects that begging is generally more common-and more lucrative-for women. The difference between Oslo and Stockholm is, however, no longer significant, suggesting that the higher likelihood of begging among those residing in Stockholm is primarily a composition effect. The lower likelihood of begging in Copenhagen remains significant, suggesting that the Danish criminal ban on begging has an independent effect on choice of strategy. Most importantly, however, the effect of Roma ethnic identity is considerably reduced. In step one, being traditional Roma increased the odds of begging with a measure of more than 8. In step two, this is reduced to approximately 2.5, suggesting that structural poverty, family social capital and oppositional identities-as measured by these five variables-go a long way in explaining why the Roma are more likely to beg.

Conclusion
The Roma of Romania and its neighbouring countries remain Europe's most marginalized ethnic group. With EU open borders and few alternative options at home, many go abroad to beg for money or engage in other types of informal street work in richer countries in the West. In many rural Roma communities, transnational migration for begging and street work has now become an institutionalized practice, constituting the backbone of the local economy. Throughout Europe, this has become a renewed source of moral panic, condemnation and ethnic prejudice towards Roma. A key purpose of this article has been to bridge the gap between overly structural and overly cultural understandings of this phenomenon. By rooting the analysis in economic sociology, I argue that we may perceive transnational migration for begging among Roma as purposeful and rational adaptations to their present situation, while at the same time being embedded in the distinct social structures and institutions characterizing Roma communities and their social position in today's Romania. Specifically, I argue that in order to understand marginal migration and begging as an institutionalized practice, three distinct factors must be taken into account.
To understand the motive for migration, one must take into account their marginalized position in Romanian society, as a result of a restructuring economy and dwindling opportunities within the Roma's traditional economic niches, the increasing salience of ethnic boundaries and discrimination in post-socialist Romania, and the severe distrust in public education which has hampered progress though mainstream institutions. To understand the ability of many Roma to engage in marginal migration despite lacking in formal resources, one must take into account social capital in the form of family and village networks, which provide people with information, opportunities, resources and support that are vital in overcoming the risks and challenges of migration outside formal institutions. Finally, to understand the willingness of many Roma to endure the stigma associated with begging one must take into account 'oppositional' Roma identities and cultural resistance towards the moral judgements of outsiders, which provide a level of psychological protection when engaging in transgressional activities. While I have tried to illustrate these separate points using qualitative data, the quantitative data provides strong support for this model's ability to account for variations in terms of propensity to engage in begging.
It should be noted that some might associate the concept of 'oppositional identities' with notions of 'culture of poverty' and moralistic 'victim-blaming' and prefer to focus on structural discrimination. However, as Brazzabeni et al. point out, a key challenge when studying the adaptations of Roma communities is to recognize their situation marked by racism, poverty and marginalization, while at the same time acknowledging some measure of agency or capacity to determine one's posture vis-à-vis their surrounding societies (Brazzabeni, Cunha, and Fotta 2015: 2). Begging may not provide a viable solution to their poverty. But given the desperate situation of many Roma communities, one may argue that going to Scandinavia to beg-without letting status-concerns or the moral sensitivities of mainstream society stop them-is a testament to their inventiveness and perseverance in a hostile world.
The policy lessons should be sobering. With free movement and open borders, states are not able to fully prevent unwanted migration from within EU. Instead, internal regulations such as anti-begging legislation are increasingly used to control the mobility of unwanted populations (Fekete 2014), and authorities in Scandinavia have more generally responded to the loss of direct control by mobilizing municipal workers and local police as everyday gatekeepers (Tervonen and Enache 2017). An implicit strategy has been to deter their entry by making life unpleasant through various forms of disruption and criminalization of their activities, while providing only a bare minimum of services. A recurring argument, for example, among local politicians in Scandinavia has been to avoid setting up public bathrooms because they might attract migrant Roma. The present analysis suggests why such efforts have only limited effect. Romanian migrant Roma are strongly motivated by a desperate situation and lack of alternatives back home; at the same time, they are quite resilient to the harassment and general unpleasantness that police, local governments and municipalities may throw at them. As long as there is money to be made through begging and street work, and the alternative options back home are worse, many will continue to brave the journey. At the same time, in order to maintain a sense of self-respect and notions of worth in the face of the contempt with which they are met both in Scandinavia and at home, they will most likely also continue to cultivate their internal bonds of kinship and the moral boundaries between themselves and the outside world.
Conflict of interest statement. None declared. Notes 1. Corte di Cassazione (Supreme Court), VI Criminal Section, No. 45516/2008. 2. RDS analysis also consists of estimators that use network size and homophily in recruitment to adjust for sample bias. However, since the analysis here will focus on relationships between variables, not population estimates, unadjusted data will be used.