Beyond the ‘Migrant Network’? Exploring Assistance Received in the Migration of Brazilians to Portugal and the Netherlands

This paper explores the tenability of three important critiques to the ‘migrant network’ approach in migration studies: (1) the narrow focus on kin and community members, which connect prospective migrants in origin countries with immigrants in the destination areas, failing to take due account of sources of assistance beyond the ‘migrant network’ like institutional or online sources; (2) that it is misleading to assume a general pattern in the role of migrant networks in migration, regardless of contexts of arrival or departure, including the scale and history of migration or the immigration regime; and (3) that ‘migrant networks’ are not equally relevant to all migrants, and that important differences may exist between labour migrants and other types of migrants like family migrants or students. Drawing on survey data on the migration of Brazilians to Portugal and the Netherlands we find support for these critiques but also reaffirm the relevance of ‘migrant networks’.


Introduction
Social network theory is well established in migration research. It has long been observed that once pioneer migration takes place, potential migrants in origin areas become connected to migration destinations, where employment opportunities and increased chances for socio-economic upgrading are anticipated. Given the access gained through social networks with previous migrants, the flow continues beyond the 'pioneers' as previous migrants mediate the migration of newcomers (De Haas 2010). Migration scholars have widely acknowledged the importance of migrant networks for reducing all kinds of social, economic and emotional costs for new migrants (Somerville 2011) and their important role in reducing selectivity and perpetuating migration (Massey et al. 1993;Massey et al. 2005;Gurak and Caces 1992;Fawcett 1989;Tsuda 1999;Pellegrino 2004). Migrant networks provide insightful information about the new society and assistance for example with housing, employment or obtaining documents, that help migrants decide and plan their move. In other words, migrant networks provide valuable forms of social capital to new migrants (Ryan 2011).
However, scholarly thought on the role of networks in the growth of international migration is also strongly influenced by Massey and colleagues' study of Mexican migration to the USA Massey et al. 1993;Massey et al. 2005;Massey and García España 1987) and there is a lack of empirical studies dealing with other geographies of migration and contexts as well as contemporary dynamics looking at different migrant profiles and the role of the internet or social media. Indeed, critiques on the initial theorisation of the role of social networks in migration have also emerged. In this article, we critically examine three of these critiques: (1) the role of other sources of assistance that migrants receive with their migration beyond a strict conceptualisation of 'migrant network', (2) the influence of contexts of arrival in the configuration of sources of assistance received with migration, and (3) the differences in the sources of assistance received with migration by different types of migration (e.g. students, labour migrants, family migrants). The article draws on survey data from a case study of Brazilian migration to Portugal and The Netherlands. Whereas we find some support for all three critiques, the relevance of 'migrant networks' is also reaffirmed. Moreover, some of the dimensions of the relevance of the 'migrant network', in its strict definition, are unpacked and situations where other actors become more relevant are identified.
and Rivero-Fuentes 2003, De Jong 2000. These two critiques to the migrant network approach have greatly enhanced our contextual thinking about the relevance of migrant networks. Yet, in this article, we focus on three other lines of critique that have so far received less empirical attention.
First, scholarly critique has pointed at the relevance of other facilitators of migration, such as employers, government officials, traffickers, other migration brokers or people beyond migrants' pre-existing network, even including complete strangers (Krissman 2005;Schapendonk 2014;Dekker and Engbersen 2014). Moreover, with the rise of the internet, nowadays, potential migrants are likely to have access to more-diverse sources of information and assistance than in the past. Following this line of argument, in order to fully grasp the complexity of migration processes it is therefore crucial to further explore the role of (potential) multiple sources of assistance beyond the 'migrant network'. Migrant networks have been conceptualised as Bsets of interpersonal ties that connect migrants, former migrants, and non-migrants in origin and destination areas through ties of kinship, friendship and shared community origin.^ (Massey et al. 2005: 42) What draws attention in this conceptualisation is its emphasis on Bties of kinship, friendship and shared community origin.^Because of the migrant network approach's focus on Breciprocal obligations based on shared understandings of kinship and friendship^ (Massey and García España 1987: 737), many other relevant actors in the migration process are effectively ignored (Bagchi 2001;Krissman 2005;Raghuram 2013;Ryan 2011). Raghuram et al. (2010: 624) for example assert that the literature on migration Bsuggests that migrants tend to be channelled and contained almost exclusively within migrant networks.^And Bagchi (2001: 9) observes that a particular shortcoming of many studies entails Btheir emphasis on the role of intimate relations (e.g., to family, friends and community members) and the subsequent neglect of more impersonal ties.^Krissman (2005: 5) argues that Bmost immigration researchers have excluded a variety of actors involved in the origination and perpetuation of migratory flows from data collection, analytical assessment, theoretical construction, and/or public policy promotion.^In doing so, Bthe migrant network concept led immigration studies astray^ (Krissman 2005: 5).
There has been empirical evidence to support this line of critique. Studying Polish migrants in Germany and Italy, Elrick and Lewandowska (2008:718), for example, find that B[recruitment] agents are significant actors in migrant networks who can be regarded as perpetuators of migration within these networks^. In the USA, scholars have also noted the importance of the Bvisa and passport industry^, which is believed to have spurred clandestine migration from Brazil (Margolis 1994: 41). In addition, Krissman (2005) draws on empirical results of five studies to support his line of critique. This evidence also indicates that different actors within migrants' networks may play important roles in some domains but be insufficient in others calling for other actors to come in to facilitate migration and generating not homogeneous but rather complex configurations in the networks that assist in the migration process.
Furthermore, the call to consider sources of assistance beyond the traditional migrant networks may have become even more salient because of the rise of the internet. Nowadays, there may be a crucial role for people who supply information online but lack any previous connection to the prospective migrants. Through the internet, those seeking to migrate establish contact with unknown immigrants in the destination country of their choice or gain access to information that they display online through YouTube, blogs, Facebook or other online forums. Qualitative studies have found that the role of the internet and social media is not restricted to assisting community life among migrants at the destination (Oosterbaan 2010;Schrooten 2012) but rather can encourage (Thulin and Vilhelmson 2014) and facilitate migration itself (Dekker and Engbersen 2014), a dimension which very few studies have examined. However, the few quantitative studies that do exist do not necessarily provide supporting evidence. Studying Indian students in Canada, Somerville and Walsworth (2015) for example, find that only 4% of the migrants they studied actually used websites in the host country as a source of information. In fact, it was the least used source of information among the study's participants. Instead, new migrants overwhelmingly relied on family or friends living abroad (Somerville and Walsworth 2015). So whereas qualitative studies have yielded high expectations regarding the role of the internet and social media, quantitative studies are scarce and suggest that its role may still be very limited. Based on the current state of research, we question (1) what actors are important sources of assistance in contemporary migration processes?
A second line of critique can be drawn from studies that have shown that contexts of destination and departure also influence migrants' use of available networks (Villarrubia-Mendoza 2016; Zell and Skop 2011;Côté et al. 2015;Gold 2001;Bagchi 2001). It is therefore misleading to assume a general pattern in the role of migrant networks in migration, regardless of contexts of arrival or departure. In our case study, we focus on the influence of destination contexts in the use of and actors involved in the networks migrants draw on to migrate. With the emergence of new destinationsbeyond long-established migration routes-it is worth examining whether different contexts of destination lead to differences in the sources of assistance and the extent of assistance sought. Massey et al. (2005) hypothesise that once migration has begun migrationsupporting institutions tend to arise to satisfy the increasing demand that follows. They see such institutions Bas a structural complement to migrant networks^ (Massey et al. 2005: 44) that contributes to the perpetuation of international migration. Following this line of reasoning, the gradual build-up of institutions that support migration is not independent from migrant support networks but an eventual by-product, or externality (Garip and Asad 2012: 6) originating from them. It is expected then that such sources of assistance beyond the migrant network (institutions) are more prominent in longstanding migration flows, where the scale of migration has generated the development of such institutionalised resources (Garip and Asad 2012: 6), than in relatively new migration flows. This leads us to question (2) Can differences be observed between two distinct destination contexts in terms of the sources of assistance provided to migrants? And if so, what kind of differences?
Third, the composition of migration flows has also become increasingly diverseconsisting not only of labour migrants but of other types of migrants as well (King 2002). The important role of 'migrant networks' had largely been established in research on labour migration flows, but it remains to be seen to what extent 'migrant networks' are also important for understanding other types of migration and, if so, in what way. Collyer (2005), for example, shows that social network theory cannot explain migration flows of asylum seekers. Furthermore, research suggests that there is a difference in the types of networks used by different occupational classes (Gold 2001). Some argue that the highly skilled rely less on migrant networks than unskilled or undocumented migrants (Zell and Skop 2011), whereas others asserts that undocumented migrants rely less on family networks due to migration restrictions (Collyer 2005). Furthermore, despite the rapid growth of the share of students in migration flows, the study of international student migration is a relatively neglected field of study (King and Raghuram 2013), calling for more research to be done on the question of how students migrate and receive assistance (Raghuram 2013). Furthermore, Bagchi (2001) asserts that a fundamental question is how the use of social networks differs between labour migrants and family migrants. In addition, van Meeteren et al. (2009) show that depending on their motivations (aspirations), undocumented immigrants mobilise social capital from different types of sources. Thus, migrants with different motivations may turn to different sources of assistance with their migration projects. This leads to our final question: (3) do migrants with different migration motives rely on different sources of assistance in their migration?
To explore these questions, we examine a larger variety of sources of assistance than is usually considered in migration research, both at the level of individuals and organisations as well as online sources-located in both the destination and the origin country-that can provide various types of assistance with the migration process.

Brazilian Migration to Portugal and the Netherlands
In this paper, we draw on the case of Brazilian migration to two destination countries: Portugal (a flow with a long history) and the Netherlands (where it constitutes a relatively recent phenomenon). We argue that this is a good empirical base to discuss these questions because both flows are heterogeneous, include different types of migrants, and are contemporary but refer to different contexts of arrival and histories of migration. In addition, in Brazilian migration, the importance of migrant networks has been identified, for example, in the case of Brazilians going to the USA (Fusco 2002;Margolis 1994), Spain (Solé et al. 2011), Japan (Zell and Skop 2011) and Portugal (Padilla 2006). Research on Brazilian migration to the Netherlands hardly exists (Roggeveen and Sandoval 2008).
In Portugal, Brazilian migration is inscribed in the historical relationships that exist between the two countries due to their shared colonial history and common language. However, before the 1980s, immigration from Brazil was not particularly significant and constituted mainly of political expatriates, married women and executives. During the 1980s, the flow grew and was mainly comprised of (i) skilled and highly skilled professionals and (ii) descendants of previous Portuguese emigrants in Brazil (Malheiros 2007).
At the end of the 1990s, a new flow emerged that became known as the 'second wave of Brazilian immigration' (Casa do Brasil/Acime 2004). This flow was large and set the basis for the constitution of the largest group of foreign nationals in the country since 2007 (data from SEF-Serviço de Estrangeiros e Fronteiras, the Aliens and Borders Service). This 'second wave' consisted mostly of workers of intermediate levels of education who found employment in low-skilled occupations in the hotel and catering sector, construction and domestic work. In recent years, the number of students in higher education has also risen considerably, from 2204 in the school year 2006/ 2007 to 8838 in 2012/2013 (Oliveira et al. 2015: 45), a trend that has been maintained.
Since the financial crisis of 2008, employment opportunities have declined, with a consequent deceleration of the migration flow to Portugal that had been motivated by economic reasons (from 32,751 new-resident permits issued in 2008 to 11,715 in 2012 and 5560 in 2014, according to SEF). During this period, there was an important flow of return to Brazil and Portuguese emigration to that country (Padilla et al. 2015: 32). A new trend has however been emerging following the impeachment of President Dilma Rousseff in 2016 and increasing political and economic instability in Brazil.
Brazilian immigration to Portugal has been diverse albeit motivated by and large by economic reasons, with great support from social networks and often preceded by internal migration (Peixoto et al. 2015: 3). Brazilian immigration has always included a high percentage of women who did not migrate within family reunification schemes (Padilla 2007). The 'second wave' of Brazilian immigration also included a large proportion of undocumented migrants. In 2003, the survey conducted by Casa do Brasil (Casa do Brasil/Acime 2004) found that 36% of respondents were undocumented. However, Portuguese governments have adopted different legislative measures that allowed the regularisation of irregular migrants throughout the 2000s: in 2001, a new legal regime (Decree-Law no. 4/2001, of 10 January) introduced the authorisation for permanence, for immigrants who were undocumented but fulfilled the requirements of the legislation, most importantly holding an employment contract; in 2004, art. 71°of the Regulatory Decree 6/2004 of 26th April instituted another extraordinary mechanism for the regularisation of undocumented migrants who had entered the country before 12 March 2003 and had been working and paying social security; in 2007, the Law 23/ 2007 of 4 July also included an exceptional mechanism (art. 88°) that allowed the regularisation of undocumented workers with employment contracts and payments for social security. In addition to these programmes for all immigrants, Brazilian nationals benefited from a specific agreement between the governments of Portugal and Brazil that became known as 'Acordo Lula'/Lula Agreement, in 2003.
Less is known about Brazilian migration to the Netherlands. In the 1980s, economically driven migration started and continues to this day. Like other Latin-American migration to the Netherlands, recent migrations are partly a result of stricter immigration controls in the USA following the terrorist bombings in 2001. Since then, there has been an increase in the Latin American population in the Netherlands (Sandoval 2008). Numbers from Statistics Netherlands show that the Brazilian population residing in the Netherlands grew from 3933 immigrants in 1996 up to 11,929 in 2012. In 2012, 68% of all Brazilians were female. However, irregular immigrants are not included in these official statistics. Estimates state that there are an additional 3000 to 20,000 Brazilian migrants in the Netherlands, most of them residing in Amsterdam ( ).

Data and Methods
This paper draws on data collected for the research project Theorising the Evolution of European Migration Systems (THEMIS). This involved a survey of 400 Brazilians residing in the Lisbon Metropolitan Area in Portugal and 206 in Amsterdam in the Netherlands in 2012.
Respondents for the survey in Amsterdam and Lisbon were found using respondentdriven sampling (RDS). This method has been used mostly with hidden populations, including immigrants, when no sampling frames are available. It is essentially a form of snowball sampling, where respondents recruit new respondents to be interviewed. Recruitment is initiated through a selected number of seeds, which were selected to be diverse in terms of their individual socio-demographic and economic characteristics, and continues thereafter through two important mechanisms: (a) incentives paid to interviewees and (b) incentives paid for the recruitment of others. The latter potentiates peer pressure to attend the interview (Johnston 2008) (see for more details Horgen Friberg and Horst 2014; Kubal et al. 2014). The objective of this method is to create long recruitment chains to the point where characteristics of the sample become independent of the original seeds. With the use of RDS, we have obtained a nonrepresentative, but varied sample of Brazilian immigrants in Amsterdam and Lisbon. The interviews were held in the language preferred by the respondents, whether Dutch, English or Portuguese.
In the survey, we asked respondents whether they received assistance with (1) obtaining information about the destination country, (2) funding travel costs, (3) obtaining documents, (4) finding a first place to stay, and (5) finding a first job. When asked about assistance with information, travel costs and documents, respondents could name a maximum of three sources. The universe of the tables we present in the subsequent sections therefore varies per type of assistance. With assistance in finding a first house and job, respondents could name only one source. In addition, we asked whether this source was located in the destination or in the origin country. Respondents' answers were then given a code using our 'code list', which specified 40 different types of individual sources (ranging from husband/wife to complete female/male stranger) and 21 institutional sources (employers, recruitment agencies including those specialised in au pair programmes and domestic work, travel agencies, migrant and other organisations, agencies that recruit or fund international student mobility, lawyers, religious institutions, embassies/consulates and migrant media as well as online sourcesin the form of websites with information). We have transformed these sources into five categories: (1) family members, (2) friends and acquaintances, (3) institutions, (4) internet sources and (5) complete strangers. Whereas the first two categories comprise sources within the 'migrant network', we consider the last three categories as sources beyond the 'migrant network'.
Following the questions introduced in the first section, we formulated the following hypotheses to be tested with the data collected: HYP 1: Sources of assistance beyond the 'migrant network' (category 3, 4, and 5) are at least as important as sources within the 'migrant network' (category 1 and 2) HYP2: Sources of assistance beyond the migrant network (category 3, 4 and 5) are more important in a longstanding migration flow (Portugal) than in a relatively new flow (the Netherlands); HYP 3: Migrants with different migration motives rely on different sources of assistance to migrate.
As hypotheses 1 and 2 are concerned with the question of how many times certain sources of assistance were mentioned by our respondents, they are approached using descriptive analyses. Hypothesis 3 states that migrants with different migration motives rely on different sources of assistance to migrate. For each source category, we have created a variable that indicates whether respondents received assistance from this source category and how many times this was indicated (max. 3). We conducted one-way analyses of variance to see if there are significant differences between migration motive categories in terms of the number of times assistance was received for each source category.
Respondents were asked to indicate whether 'experiencing the culture and life in another country', 'opportunities for work', 'opportunities for studying', 'learning a language', or 'being with family members or other people you care about' was an important motive for their move, and, in a subsequent question, which of them they considered the most important. We categorised respondents based on the latter. Table 1 shows that in both countries, the largest share of migrants are primarily driven by work motivations, the majority of them found jobs in low-skilled or semi-skilled occupations; only a minor proportion were professionals on departure or were inserted in highskilled occupations upon arrival but that the other motives are also well represented.
Through the analysis of our qualitative material (van Meeteren andPereira 2013, 2016) and by looking at the secondary motives in the survey, we found that those who indicated they came to learn a language were very similar to those who came because of opportunities to study in the Netherlands. They are similar with regard to their sociodemographic profile and the assistance they received with their migration. In addition, from the qualitative material, we gathered that the motives themselves are in fact quite similar (see also Martes 2011): they have come to learn something. We have therefore combined these two motivations into one 'study' category, but only for those who also indicated that 'study' was a secondary motive. Table 2 displays important respondent characteristics by migration motive and destination country.

Results: The Relevance of Sources of Assistance beyond the 'Migrant Network'
In this section, we first test the hypothesis that sources of assistance beyond the 'migrant network' are at least as important as sources within the 'migrant network'. We test this hypothesis on five types of assistance: assistance with obtaining information (Table 3), travel costs, documents, finding a job and finding a house. The table below describes the number of times each source category was mentioned by respondents as a source of assistance with obtaining information. Table 3 shows that 'migrant networks' are key providers of information: migrants most often relied on friends and acquaintances (first) and family (next) in the destination country to obtain information. However, with 16%, institutional sources    (7) 19 (7) Institutional 28 (9) 8 (3) 36 (12) Institutional 10 (4) 180 (67) 190 ( (60) 123 (40) 308 (100) Total 20 (7) 249 (93) 269 (100) (especially in the origin country), with 9% online sources and 2% assistance of complete strangers, the role of assistance from sources beyond the migrant network (27%) should not be neglected. Assistance with funding travel costs is largely provided by family members, in both the origin and the destination country (Table 4). Institutional sources based in the destination country are mentioned as a major source of assistance with obtaining documents (71%). In fact, with this type of assistance, they are the most relevant source of assistance in Portugal, as we shall see later on. Internet sources and strangers were not mentioned at all for these forms of assistance.
When we examine assistance received with finding a first house and a first job (Table 5), we see that friends and acquaintances were mentioned most often. Nevertheless, there is a minor role for assistance from institutional sources, internet and strangers.
Overall, sources of assistance beyond the migrant network are less important sources of assistance than sources within the 'migrant network', thus rejecting our first hypothesis. The traditional 'migrant network' still plays a major role, and is supplemented by other relevant sources. The extent to which sources beyond the 'migrant network' are important depends on the type of assistance offered. Institutions can be relevant sources of information and give access to regularisation in the destination country if the policy and legal framework grants them that role. In such cases, they are actually the most important sources of assistance received. Internet sources were not as prominent as one might expect. They are a relevant provider of information but not of assistance in the other domains. It appears that different sources in different locations (origin/destination) provide different forms of assistance.

Results: Differences Between a Long-lasting Flow and a New Migration Flow
Interesting differences are observable between the two destination contexts. Overall, migrants who moved to the Netherlands more often sought information than those migrating to Portugal, most likely because of the higher familiarity of Brazilians with Portugal, given the historical relationship between the two countries and the shared language. The only exception concerns obtaining information on how to find housing. This can be directly related to the high regulation of the Dutch housing market as well as the language barrier that constitute important impediments to find housing on one's own, particularly for undocumented migrants (van Meeteren et al. 2009; van Meeteren 2014). Because of this, migrants need to ensure they have a place to stay before arrival.
In addition, Brazilians in Portugal have made the most use of institutions as sources of assistance to obtain documents, which by and largely reflects the role played by employers and agencies that recruit international students (including universities) in assisting newly arrived immigrants to obtain the appropriate documents to remain in Portugal (the role of such institutions is minimal in the Netherlands). In this domain, assistance is more noteworthy in Portugal than in the Netherlands. In the Netherlands, Brazilians are allowed to enter as tourists with visas granted at the border, but then high numbers stay on irregularly as there are scarce opportunities to gain access to permits after that. For most, the only way out of irregularity is through marriage (van Meeteren 2014). In Portugal, access to visas and permits to remain in the country after the tourist visa expired is easier than in the Netherlands. Opportunities existed for some time in Portugal for those able to present a valid working contract and respective social security and tax payments, as was mentioned before. Therefore, employers in Portugal are most frequently cited by respondents as a source of assistance with obtaining documents to remain in the country. Table 7 shows us that there are no significant differences observable between Portugal and the Netherlands when it comes to the sources of assistance with finding   (100) 178 (100) a first house and a first job. In both countries, family and friends are the most cited sources of assistance. Reviewing the information in Tables 6 and 7, we conclude that our second hypothesis-that sources of assistance beyond the migrant network are more important in Portugal than in the Netherlands-can be confirmed only for assistance with obtaining information, travel costs and documents. However, it probably does not have anything to do with the maturity of the flow. In Portugal, institutions emanating from the larger scale and longer period of the migration flow there did not seem very relevant as providers of assistance as predicted by the 'externalities' mechanism identified by Garip and Asad (2012: 6) or the importance of the 'maturity' of the flow proposed by Massey et al. (2005). Where institutions did play a role, they have been mostly independent from the critical mass of the migration dynamics. Instead, this role was related to the policy context, as previously explained.

Results: The Differentiating Role of Migration Motives
In this section, we assess our third hypothesis: migrants with different migration motives differ with regard to the sources of assistance they receive with their migration.
The analyses (Table 8) reveal a significant relationship between migration motives and all sources of assistance with information. If we look a bit more closely and examine which pairwise comparisons render significant differences, we see that respondents who indicated their primary motive was to join family or loved ones Each motivation has been assigned a small letter ranging from a to d (see first row). The small letters in the table indicate the reference category in relation to which a significant result was found. Hence, in the column 'study', the result in the first row indicates that migrants with a study motive significantly differ from migrants with a family motive in the assistance they received with obtaining information *The result was significant using a one-way ANOVA. However, as the data are not normally distributed and equal variances could mostly not be assumed, we have performed non-parametric tests (with Kruskal-Wallis replacing the ANOVA and with Mann-Whitney tests checking the post hoc ANOVA results). The results were generally the same. If one test gave a non-significant result but the other gave a significant result, we also report the result of the non-parametric test between brackets significantly differ from the other categories of motives: they received information from family members significantly more often than immigrants who came to experience life and culture in another country (p < .00), who came to work (p < .00), and who came to study (p < .00). If we examine information received from family in the origin and destination separately, we also see that workers received significantly more information from family in the destination country than students did (p > .05). If we examine pairwise comparisons for information received from friends and acquaintances, family migrants again stand out: they received significantly less information from friends and acquaintances than migrants with other motives. Students stand out regarding the information that they have received from institutions. Workers mentioned internet sources significantly less often as a source of information than Each motivation has been assigned a small letter ranging from a to d (see first row). The small letters in the table indicate the reference category in relation to which a significant result was found. Hence, in the column 'study', the result in the first row indicates that migrants with a study motive significantly differ from migrants with a family motive in the assistance they received with obtaining information Each motivation has been assigned a small letter ranging from a to d (see first row). The small letters in the table indicate the reference category in relation to which a significant result was found. Hence, in the column 'study', the result in the first row indicates that migrants with a study motive significantly differ from migrants with a family motive in the assistance they received with obtaining information migrants with other migration motives. Overall, this analysis shows that migrants' primary migration motives are an important differentiating factor when it comes to understanding how information is provided in migration processes. In other words, the sources of assistance used for obtaining information differ by motive of migration. When we examine assistance received to fund travel costs (Table 9), we see that students have received significantly more assistance from institutions than from family members. Furthermore, the relationship between migration motive and assistance received from family only becomes significant if we examine the origin and destination country separately. This is because family migrants have received significantly less assistance with travel costs from family members in the origin country, while they have received significantly more assistance with travel costs from family members in the destination than migrants with other migration motives. This shows that it is crucial to consider also the place where the source of assistance is located and that not all migrants depend to the same extent on previous migrants to fund their migration.
We have seen in the previous section that very few migrants received assistance with documents from friends or acquaintances. They generally mainly received assistance with documents from family or institutions. However, migrants with different motives differ in the extent to which they have received assistance from one or the other (Table 10). Family migrants stand out compared to students and workers. While they received significantly more assistance with documents from family members than students and workers, they received significantly less assistance from institutions than the others.
Almost half of our respondents came to live with someone who already had a house. Nevertheless, there are significant differences between respondents with different migration motives; F(3,592) = 14.95, p < .00. In fact, migrants driven by family motives stand out because they came to live with someone significantly more often than those with other migration motives (p < .00).
For only a few Brazilians (2%), housing was arranged in connection with employment or studies. Those who did not come to live with someone and for whom housing was not arranged often received assistance with finding a first place to stay (39%). However, there are significant differences between migrants with different migration motives (Table 11). If we examine Table 11, we will see that there is a small significant association between the migration motive and the source category from which support with finding a first house was received, with Cramer's V(6) = .28, p < .00. We see that migrants who came to experience life and culture in another country, workers and students mostly received support from friends and acquaintances with finding a house, whereas family migrants mostly received support from family members. In addition, there is a small assistance role for institutions for all motives except for family migrants. Of all the respondents who work or have worked in the destination country, 80% received assistance with finding their first job. Table 12 shows that there is a small but significant association between the migration motive and the source category, with Cramer's V(6) = .15, p < .00. However, regardless of their migration motive, many Brazilians received help from friends and acquaintances to find their first job (see Giulietti et al. 2013 for similar observations).
We see that assistance from friends and acquaintances plays a crucial role for all migration motives in both destination countries in finding a first job. This is likely to reflect the specificity of this type of assistance. Studies have indeed noted the significance of weak ties in finding employment (Granovetter 1973), particularly for undocumented migrants (van Meeteren et al. 2009; van Meeteren 2014). Furthermore, family members mediate the contact of the migrant with others who are better positioned in the labour market them a job.
Our third hypothesis was that migrants with different migration motives differ with regard to the source they receive assistance from with their migration. Our results indicate that this can be confirmed for most types of assistance, both with regard to the type of source and the place where the source is located. In most cases, family migrants differentiate from the rest because they tend to rely on family (located mostly in the destination country) more than the others. In addition, students also stand out for the assistance received from institutions (with information and funding). Other interesting results are the key role of friends and acquaintances to enter the labour market (also for family migrants) and that workers rely less on internet sources to get information than the others.

Conclusion
We have scrutinised three important critiques to the dominant 'migrant network' approach to the study of migration and found some support for all of them. At the same time, we also confirm that migrant networks are still very important to understand migration flows-more so for some forms of assistance than others. Migrant networks constitute an important resource for new migrants; however, it is not the only resource that they use. Thus, the theorisation of migration processes and mechanisms of assistance need to be much more nuanced to include sources of assistance beyond the 'migrant network', consider context-specific circumstances, and also migration motives. In this sense, the critique of scholars that the 'migrant network' approach implies a too narrow research focus (Bagchi 2001;Krissman 2005;Raghuram et al. 2010;Ryan 2011) has been warranted by our findings-yet the importance of sources of assistance beyond the migrant network may be more limited than some may have expected. Nevertheless, our findings demonstrate that migration scholars need to move beyond the once common interpretation of the fundamental role of social networks as static 'migrant networks' based on pre-existing community or kinship relationships. Instead, we need to consider that (i) a large share of migrants rely on funding resources that are available within their families in the origin country and therefore disconnected from previous migration dynamics and (ii) a broad range of providers of assistance are involved in migration processes (in both the origin and destination country, personal and institutional, stemming from pre-existing ties and new ones) for different types of assistance, for migrants with different migration motives and for different contexts of reception.
Taking into account, these three critical dimensions also allows us to challenge the idea that a pre-existent connection to previous migrants is essential for further migration, or that the 'migrant network' remains the critical explaining mechanism shaping the perpetuation of migration. We propose that existing resources, not only inside but also outside one's networks, create multiple avenues that open up migration possibilities. For example, we have found that institutions outside the migration dynamics-such as employers, student recruitment agents or higher education international departments-constitute important sources of information, assistance with regularisation and funding for some migrants (mostly workers and students). As such, we provide a more nuanced and detailed analysis of the sources of assistance involved in the different dimensions or stages of the migration process for different types of migrants. This is a step further towards the understanding of migration as a social process, which is crucial for anticipating future flows (Garip and Asad 2012:2). Schapendonk (2014: 2) notes that networks are often conceived of in migration research as static entities that link pre-existing ties, suggesting that Bnetworks are simply 'out there' as a kind of social given, and that migrants only have to plug in to these networks^to gain access to the information and assistance needed to put in place the migration project. However, social networks do not automatically lead to a transfer of assistance. The social capital potentially available in such networks needs to be mobilised through networking strategies. Indeed, most migration research dealing with the role of social networks as facilitators of migration have failed to look at how and for what purpose migrants mobilise different members within their networks. Thus, there has been little detailed account of which agents are engaged in the different dimensions or stages of the migration process (preparation, funding or assistance on arrival). For example, our analysis has clearly shown that while family members are crucial providers of assistance with funding and housing, they are less relevant when it comes to entering the labour market, a domain where friends and acquaintances constitute more relevant sources of help. In addition, a static conceptualisation of social networks implies that researchers tend to overlook other sources of assistance for migrants, for example, online sources or institutional sources to get information.
All in all, new theoretical developments are needed if we are to fully understand the mechanisms that enable contemporary migration flows. If anything, our paper clearly shows the need for those theoretical developments not to be on a general level and applied to all types of migration and contexts of reception. Those theoretical advancements need to be contextualised, differentiating between different sources in distinct locations that provide multiple forms of assistance to migrants with diverse migration motives.