Migration, Development, Gender and the ‘Black Box’ of Remittances: Comparative Findings from Albania and Ecuador

Set within the growing literature on migration and development, this paper has two interlinked objectives. First, it examines remittances, a key element of the migration-development nexus, from a gendered perspective. Second it does so in a comparative empirical perspective, focusing on remittance behaviour in two contrasting settings, albania and ecuador. both countries have experienced mass emigration in recent decades. research is based on household surveys with remittance receivers in selected rural areas of both countries, supplemented by in-depth interviews with both senders and receivers of remittances. by using the concept of ‘remittance dyads’ – person-to-person transfers of money and gifts – we examine the gendered mechanics of conveying and managing remittances to see if they have the potential to reshape gender relations in these migrant households. they do, but the effects are limited. Albania and Ecuador; and an outline of the survey methods used to collect primary data. In the ‘results’ sections of the paper we examine the dyadic patterns of remittances from New York to Ecuador and from Greece to Albania. We explain how these reflect the male-led character of the two migration systems, and we evaluate the extent to which evolving remittance dynamics have the potential to re-texture gender relations. In the two case-studies under review we find that changes in gender relations are rather modest.


Introduction
There has been a surge of interest over the past decade in the relationship between migration and development. The def ining contribution which kick-started this trend was Van Hear and Sørensen's (2003) The Migration-Development Nexus, wherein remittances were highlighted as the key contribution that migration could make to the improvement in well-being of people in migrant source countries. Indeed, remittances development? And if so, does this development feed further migration, as postulated above? Or does migration act to further reproduce the condition of underdevelopment? And if so, does continuing underdevelopment lead to continuing migration? In other words, does the migration-development nexus take the form of a virtuous cycle, or a vicious one?
2.1. Optimism, pessimism, and pendulum swings As Hein de Haas has recently argued (2010,2012; see also Faist and Fauser 2011), the debate between these two opposing interpretations has swung like a pendulum, from optimism in the 1950s and 1960s to pessimism in the 1970s and 1980s, and back again to optimism in the 1990s and 2000s. Now, we detect the start of a new swing towards pessimism.
The industrial expansion of Western Europe and North America in the early postwar decades was sustained by large-scale worker migration from Southern Europe, Latin America, and elsewhere. Neoclassical economics 'explained' this mass transfer as a spatial reallocation from labour-rich to labour-poor countries, spurred by the market incentive of much higher wages. Under this model, migration continues until there is factor-price equalisation, at which point migration stops because wage discrepancies are no greater than the cost of migration. Also in the 'pure' neoclassical model, there are no remittances and no return migration, since migrants are viewed as having moved for good as part of the adjustment to spatial equilibrium of labour. However, in many European countries, migrants were treated as 'guest-workers' who would eventually return home. Part of this migration management regime was the 'developmentalist' assumption that migrants would take back savings, training, new ideas and entrepreneurialism; all of which would stimulate development in their home countries. For the most part, these were pious hopes. Detailed studies of Turkey, the country which had supplied the largest number of migrant workers to Europe, questioned both the scale and nature of the 'emigration of surplus labour and return of innovation' model (Abadan-Unat et al. 1974;Paine 1974). The swing to pessimism had begun.
Incorporating late-1960s' thinking from the Latin American 'dependency school' (Frank 1969) as well as Third-Worldist concerns over brain drain from Africa and Asia (Bhaghwati 1976), growing pessimism over the economic and moral value of migration reflected Marxist theories of capitalism's exploitation of (migrant) workers. Applying the dependency model to Europe, Seers (1979) argued that periphery-to-core migration was the syndrome of the Frankian notion of the 'development of underdevelopment' rather than a stimulus to development in migrant-origin countries. Remittances -hardly noticed in the neoclassic model -were acknowledged but they were not seen as a productive input into the development process. Rather, it was alleged, they were 'frittered away' on conspicuous consumption, used to build unnecessarily luxurious housing, or spent on other 'non-productive' outlets such as small shops in migrants' home villages (Lipton 1980;Rhoades 1978). According to Castles and Kosack, in a book which captured the theoretical zeitgeist of the time, 'labour migration is a form of development aid given by the poor countries to the rich ones ' (1973: 8).
Around 1990, the pendulum swung again, back to what de Haas (2012: 19) calls 'neo-optimism'. Drawing partly on de Haas (2010Haas ( , 2012 and Faist and Fauser (2011), we posit four main reasons for this. First, the 1980s and 1990s saw a body of empirical research which suggested that, in certain circumstances, migration could stimulate home-country development. In other words, alongside negative 'backwash' effects (depopulation, brain drain, inflation etc.) were also 'trickle-down' effects whereby remittances and migrants' investments gave a multiplier boost to the local economy.
Second, the neo-liberal, individualist ethos of the 1990s and 2000s placed migrants centre-stage as development actors, in contrast to the neoclassical and neo-Marxist ideologies which viewed migrants either as atoms responding to market forces or as victim-pawns of the exploitative capitalist system. Alongside the celebration of the migrant as the 'hero' of development comes an emphasis on remittances: not only financial remittances but also backflows of knowledge, ideas and new behaviours -what Peggy Levitt (1998) called 'social remittances'.
Third, migration-development neo-optimism was underpinned by the 'new economics of labour migration' (see Lucas and Stark 1985;Stark 1991;Stark and Bloom 1985). Although NELM drew on some aspects of the neoclassical model, it incorporated two fundamental differences from the earlier way of thinking. First, it aggregated migration decision-making and economic behaviour to the 'meso-scale' of the family and community; and second it combined income maximisation with risk aversion. In this way, migration and remittances were seen as a hedge against the failure of other elements of household income, such as a crop wipe-out. In short, NELM reconceptualised migration as a household livelihood strategy, with remittances as the central plank.
The final paradigm shift which gave theoretical purchase to the new optimism was the transnationalist view of migration introduced by Glick Schiller et al. (1995). This, too, was a migrant-centred model, which saw migrants as agents of their own development, and that of their home communities. The 'transnational turn' recognised the increased possibilities of migrants and their families to live more or less simultaneously in two (or more) places, thanks to improved techniques and lowered costs of international travel and communication. Remittances are seen as the key element of the economics of transnational living (Guarnizo 2003).
Especially over the past decade, the 'virtuous' interpretation of the migration-development cycle has dominated international discourse, including the Global Forum for Migration and Development, the World Bank, UNDP, and the UK's Department for International Development (Piper 2009). The consensus view is that migration can be a 'route out of poverty' for many and a pathway to prosperity for some; therefore, echoing the title of the 2009 UNDP Human Development Report, the barriers to mobility need to be overcome. Richard Black, director of a major DfID-funded programme of research on migration and poverty at the University of Sussex, argues that migration should be considered an international 'public good' because of its potential to alleviate poverty and stimulate development (2011).
More sceptical views emerged at the turn of the decade. Delgado Wise and Márquez Covarrubias (2011) restated the neo-Marxist critique, basing their arguments on their interpretation of the 'asymmetric and subordinated' Mexico-US migratory system, with migrant workers incorporated into the lowest and most exploited segments of the highly polarised US labour market. Meanwhile Skeldon (2008) warned of the danger of expecting too much from migrants' capability to effect real development, and asked whether it is morally justified to impose on migrants a 'duty' to stimulate home-country development when their lives -thanks partly to the global regime of migration control which forces them into irregularity -are blighted by low wages, precariousness, deskilling and racism (Bakewell 2007). Hence, 'we may be at a new turning-point… heading towards a neo-pessimistic backswing of the migration and development pendulum' (de Haas 2012: 22).

2.2.
Opening up and gendering the 'black box' of remittances Throughout the above debate on the shifting terrain of the migrationdevelopment relationship, remittances have held a fluctuating and often vicarious position, coming increasingly to the fore in the neo-optimism phase which still characterises present-day policy. Economists and development planners have latched on to remittances as a financial flow to developing and transition countries which has to be maximised, and prudently invested, without paying much attention to their ontology and phenomenology. What does the sending and receiving of remittances mean to the individual actors involved? Who, exactly, sends; and who receives? How do they define what 'counts' as remittances and what does not? In short, we see remittances as a 'black box' that needs to be opened up and peered into in order to understand what is really going on.
Building on the now-well-established transnational paradigm, we conceptualise remittances as transnational gendered social practice which reflects the gender structures of both the societies of migrant origin and destination, as well as the gender relations which are embedded in the migration and remittances processes themselves. Unpacking the gendered social practice of remittances, we observe a variety of 'remittance dyads' (Carling 2007). The remittance dyad is an individual sender-to-receiver pairing, but variety is introduced by the distinction between primary and secondary dyads. A male migrant could have as his main dyadic partner his wife in the home village, but might also remit smaller amounts to his parents, his secondary dyad, perhaps privileging his father as the recipient. Thus we begin to see the relevance of a gendered analysis of remittances, which is as yet underdeveloped.
The existing literature on gender and remittances falls into two camps. First there are papers which treat gender purely as a dichotomous variable in looking at whether men or women are the more reliable remitters, and who remits more. Reviewing this literature, Carling (2008b: 588) finds it inconclusive. It often seems that men are more likely to remit, and to remit larger amounts, but women may remit a larger proportion of their (generally lower) wages than men. Another common view is that women, because of their stronger family orientation and nurturing nature, are more 'sensible' receivers and managers of remittances than men; women direct remittances to family welfare, health and education, men towards land, housing, vehicles, and gadgets (Rahman and Fee 2009). However, there is little systematic evidence, let alone statistical proof, to support these generalisations.
The second strand of literature challenges the 'Are men or women better remitters?' discourse and tries to unravel the gender relations behind the entire migration-remittances cycle (King and Vullnetari 2010). A landmark in this more integrated approach was the programme of research launched in the latter half of the 2000s on 'Gender, Remittances and Development' by the United Nations International Research and Training Institute for the Advancement of Women (UN-INSRAW). In their working paper which acted as the ideological and literature-survey platform for the UN-INSTRAW research, Ramírez et al. (2005) make two key points: -Gender is a cross-cutting element throughout all stages of the migration process. It not only influences physical movement across national borders, but gender relations are themselves negotiated and either reconfirmed or reconfigured within transnational migratory contexts. -Remittances are more than just periodic financial transfers: they are the result of complex processes of negotiation within households that are immersed in intricate networks of relations between the diaspora and the countries of origin. Above all, decisions about how remittances are spent, who benefits, and longer-term effects on family welfare and structure, are not gender-neutral.
These orientational statements prompt further consideration. We can deepen the theoretical analysis by placing it with the 'gendered geographies of power' framework of Mahler and Pessar (2001). The three key components of this conceptual model are, firstly, the notion of power geometries. As Doreen Massey points out (1994: 149), flows of people, goods, money and images in our highly globalised but unequal world are subject to different constellations of power. Secondly, social location refers to individuals' position within power hierarchies, which include social-class and kinship structures. Thirdly, geographic scale captures the way in which genderpower relations operate across multiple levels, from the (female) body which in various ways is controlled or excluded, to the meso-scale domains of family and community, to the national and international scales. We now move to the empirical part of our paper in which we attempt to answer questions relating to the foregoing conceptual outline at two levels. First, in relation to our scoping of the migration-development nexus, do the cases of Albania and Ecuador broadly fit the virtuous or the vicious cycle view of the nexus? Second, focusing more specifically on remittances and their link to the migration-development debate, how are these financial and allied transfers gendered in our two case-studies? This, in turn, breaks down into two recursive questions: -What are the gender and family structures that shape migration and remittance behaviour? -How does the sending and receiving of remittances reshape gender relations in migrant communities and sending areas?

The settings: Albania and Greece, Ecuador and the United States
The research is based on two international migrations: one short-distance, linking adjacent countries in the Balkans, the other a globe-spanning South to North 'hemispheric' migration. The Albania-Greece study was carried out by King and Vullnetari as part of the aforementioned UN-INSTRAW programme of research (see Vullnetari and King 2011;also King et al. 2011). The Ecuador-US study is drawn from Mata-Codesal's doctoral thesis (Mata-Codesal 2011). Table 1 sets out some comparative statistics on the two migrant-origin countries. Several figures are strikingly similar: human development, life expectancy, GNI per capita, recent GDP growth, stock of emigrants, and the share of emigrants who are tertiary-educated. Others are different: the relative scale of emigration is much greater for Albania, likewise the weight of remittances per head of the country-of-origin population. The average remittance sent per emigrant is higher for Ecuador, a fact that reflects the greater income divide between Ecuador and the US on the one hand, and Albania and Greece on the other. Albanian emigration started in 1991, following 45 years of harsh communist rule during which emigration was banned. Whilst part of the eagerness to migrate was the natural curiosity of people to see a world that had been denied them, the most important push factor was the dire economic situation produced by the chaotic transition to 'democracy' and a free-market economy, which left many people without jobs or other means of support. By 2010, 1.4 million Albanians were abroad, most of them in Greece (600,000) and Italy (400,000). Throughout the early years of emigration, most of the movement was irregular, since Albanians had few legal opportunities to enter and legally work in Greece, or elsewhere. Entry to Greece was on foot, using paths which crossed the mountainous border, dodging the Greek border police. Given the nature of these treks as arduous and dangerous, as well as the patriarchal nature of Albanian families which limited the independent mobility of women, most of the emigrants in the 1990s were men.
A major change in the structural position of Albanians in Greek society occurred in 1998 when a regularisation scheme was launched: more than 200,000 Albanians legalised their presence in Greece. Further regularisations took place in the 2000s. These measures allowed Albanians to improve their access to the labour market and protect themselves from the kinds of exploitation that 'illegal' immigrants are vulnerable to. Regularisation also gave them the basis to bring over their wives and families, or to start a family in Greece. This combined legal, economic and family-structure transition is crucial for understanding differentiated patterns of remittance dyads. Men have continued their initial occupational specialisations in the farm-labour and building sectors but have been able to access (at least up to the recent severe recession in Greece) more secure work, including setting up their own small businesses in fields such as construction, removals, house repairs etc. Women work mainly as domestic cleaners or as carers of small children and elderly people (Hatziprokopiou 2003).
Emigration from Ecuador to the US dates to the 1950s, but accelerated during the 1990s due to economic stress produced by structural adjustment measures (Pribilsky 2007). Towards the end of that decade the economy collapsed and in 2000, in a desperate measure to curb hyperinflation, the US dollar replaced the Ecuadorian sucre as the country's currency. This monetary switch proved socially regressive, ruining the already-precarious financial situation of the poor and less well-off. The political situation was equally unstable, with ten different governments during 1996-2007. Faced with this dire economic and political situation, many Ecuadorian men emigrated in order to earn remittances to support their families. Even though the US became progressively more expensive and risky compared to European destinations (notably Spain), Ecuadorians continued to head for New York since the monetary rewards are perceived to be higher there. 1 Recent estimates give figures of approximately 500,000 Ecuadorians living in the US, and 400,000 in Spain (Gratton 2007). Emigration to New York is male-dominated and irregular: most men work in construction or in other manual occupations where employers ask no questions about legal status. Emigration to Spain is female-dominated and has fewer obstacles: most entered as visa-free tourists and overstayed, subsequently taking advantage of Spain's periodic regularisations. 2

Research design and methods
Within the two bi-national contexts, field research was articulated along localised 'migration-remittance corridors'. Three contiguous villages in the municipality of Burim (a pseudonym) in southern Albania were chosen to represent the migration trends and ethno-religious characteristics of south-central Albania (Muslim majority, with some Orthodox Christians and socially marginalised Roma). The villages have a combined population of around 7000 and are located at some distance from Korçë, the most important administrative and economic centre in this part of Albania. Burim lies at an altitude of 800 metres within an elevated plain with favourable soils for farming. Summers are hot and dry, winters wet and cool with occasional snowfalls. Mixed agriculture is practised -cereals, vegetables and fruit (apples are a local specialism), plus pigs, cattle, sheep and poultry. Although there are some signs of a remittance-led revival, agriculture has languished into semi-subsistency due to the break-up of the communist-era cooperatives into smallholdings and the abandonment of infrastructures, as well as the effects of emigration itself. An estimated 80 per cent of emigration from Burim is to Greece, most of it to Thessaloniki, the second city of Greece, more accessible than Athens. Xarbán, the Ecuadorian field location (also a pseudonym), has a population of just over 2000; it is in the southern province of Azuay, whose capital Cuenca is about an hour and a half away by bus. Xarbán is positioned on the lower slope of a large valley system. The lowest land, at around 2500 metres, is given over to agriculture (corn, potatoes, tomatoes, beans, fruits); the middle-range land is cattle pasture; the highest land, above 3500 m, is unusable. The climate is high-altitude tropical, with abundant rain but also a dry season. Mounting population pressure on hilly and erodible land has progressively split up ownership into minifundios (less than 5 hectares) and microfundios (< 1 ha). Whilst agriculture provides the basis for survival, it allows for little improvement in living standards or life ambitions (Carpio Benalcázar 1992: 46). Migration, initially internal (to the coastal sugar estates and the Amazon basin) and then international, to the US, became the favoured way to progress. Emigration became large-scale during the 1990s and 2000s following a destructive landslide in 1993 and the economic problems referred to earlier. As with Albanian migration to Greece, the recession years since 2008 have had a dampening effect on emigration. The much smaller number of female immigrants in New York are involved in raising families and working part-time as domestic cleaners and baby-sitters. Given the cost of migration, on average $15,000 per head, most migrants can only start to rend remittances after a year or two, once their debt is paid off.
Both projects comprised two main field instruments: a village-based household survey to remittance receivers; and follow-up interviews to some remittance receivers, and to remittance senders in New York and Thessaloniki. Although the sample sizes of the questionnaire surveys differed (350 households in Burim, 76 in Xarbán), the sample fraction of the village populations was roughly equal at around 15 per cent. 3 Whilst the two investigations were independent, the survey schedule contained many questions on remittances which were identical or closely similar -notably those relating to dyads, amounts sent, frequency and means of transfer, management of remittance spending, and the uses to which remittances were put. Both surveys were administered via a process of random and snowball sampling. We are confident that, whilst statistically 'true' representativeness cannot be guaranteed, there is a good correspondence between the samples taken and the wider population of village households with migrants abroad.
In both Burim and Xarbán, 25 in-depth interviews were carried out in the respective communities with remittance receivers, local key informants and returnees. A further 20 interviews were taken with remittance senders in Thessaloniki, and 10 in New York. Where possible, interviews were recorded and transcribed; otherwise extensive notes were taken. The formal research instruments were supplemented by ethnographic observations in all research sites.
A final note on timing. The fieldwork was carried out between late 2007 and early 2009; hence before or during the early months of the global recession, which for that reason does not feature much in our discussion.

Sending and receiving remittances
We start with background data from the household surveys (Tables 2 and 3). For remittance amounts, comparability is slightly compromised by the different currencies used. However the mean amounts per year -€2600 for Burim and $3430 for Xarbán -are remarkably similar given the higher value of the euro. What differs is the distribution of amounts, with Burim remittances more tightly bunched around the modal class of €1000-2000. For Xarbán, a greater share of households receives either lower or larger amounts. These differences mainly reflect the wage limitations of the labour market for immigrants in Greece. The New York economy offers greater earning opportunities, but not for all immigrants, and some may still be paying off debts to smugglers.    Tables 4 and 5 set out the main remittance dyads for Xarbán and Burim respectively. The listing of the dyads is in their order of frequency as reported in the survey, interview and key-informant data. Kinship, gender and generation are the main structuring variables for these dyads, but issues of power, particularly patriarchal power, are the real 'drivers' behind the relationships mapped out. What is also interesting -and this is one of our main research questions -is how the transnationalisation of these kinship links through migration and the experience of sending and receiving remittances acts to modify, or even reinforce, these gendered power-geometries.

Remittance dyads and gender-power relations
Both the Albanian and Ecuadorian societies can be regarded as patriarchal; however we perceive Albanian rural society as more deeply structured along patriarchal lies than the Ecuadorian one. Ecuadorian village society reflects the kind of patriarchal and matriarchal values sanctioned by the Catholic Church: a system of fairly traditional gender roles but with women retaining some agency within the maternal and caring spheres and engaging in farm labour as childrearing and other responsibilities allow. The nuclear family predominates, although kinship links remain strong (Kyle 2000;Pribilsky 2007;Weismantel 1998).
Albanian rural society conforms to many of these same gendered values, but there are essential differences. Religion, outlawed by the communist regime which declared Albania 'an atheist state', plays a minor role although people are aware of their Muslim, Orthodox or mixed heritage. What distinguishes the Albanian case is the more hierarchically ordered gender and generational structures, in particular the 'ownership' of the daughter by her father, and the way in which this 'possession' is passed to her husband and his paternal family upon marriage (de Waal 2005; Saltmarshe 2001). These gender-power relationships, 'legitimised' by Albanian customary law and by no means eradicated by the communist regime's commitment to gender-equality, are more firmly entrenched in the north of Albania; less so in the south, where Burim is located. Our data from southern Albania show some departures from this normative patriarchal framework, as we shall see. However we start with Xarbán as the patterns here are more straightforward.

5.2.
Xarbán and New York From Xarbán emigration started as a male-led phenomenon and has remained so, due to the continued 'illegal' status of the migrants in the US, which makes family settlement very difficult. This, rather than patriarchal principles per se, has determined that remittance-senders are mainly males migration, development, gender and the 'blaCK box' oF remit tanCeS living abroad as single men, whether they are married or not. For Xarbán, by far the most common dyad comprises married men who remit to their wives (Table 4). When married migrants are still childless, the 'left-behind' wife feels more vulnerable: If the money ever stops arriving and I am used to living a comfortable life, that's bad. Suddenly something goes wrong with my husband… then I am fucked -(Gabriela, age 45, no children).
All respondents (senders and receivers alike) agree that the obligation to send remittances becomes stronger once children are born, especially if there is at least one son: Once I had my kids my life improved. Before that, he sent $20 a month to me; he didn't care whether I had shoes on my feet or not. Now he sends $400 a month. Because of the children. He must send [money to support them] -(Berta, 30, two daughters and one son).
In the minority of cases where the migrant unit is the married couple, other dyads are formed. In these cases, often the children remain in Xarbán, but the form of the dyad depends on the age of the children (Table 4). If the latter are old enough to live independently, they are remittance recipients, usually father to son. If they are younger, they live with relatives, usually with their maternal grandmother, or as a second option with a maternal aunt. Note how, with these latter dyads, female kinship patterns prevail over male ones -in contrast with the Albanian case. Also in Albania, it is rare to find the parent-offspring dyad.  Table 4 is the single unmarried male to his parents -usually to the father, even if the responsibility for managing the daily household budget falls to the wife. Given the relatively large size of families in rural Ecuador, many village-based families have more than one son abroad and life for the parents can be relatively well-supported. Sometimes, indeed, both the husband and one or more sons are in the US. As Rosa (51, husband and three sons abroad) put it, 'Here you can live well on money from the US'. But this improvement in material well-being is not without its negative side. Loneliness and long-term physical separation are often mentioned as major drawbacks of emigration by remittance receivers:

The final dyad in
Sometimes I miss them [her children abroad]. I did not want them to emigrate, but they told me: 'We can't just stay here with you, starving.' Now, I buy 50 cents of bananas and they go bad, there is no one to eat them. I wish they would be here to eat them. When people are alone, this is bad. But you have to be hard-hearted -(Graciela, 48, widow with four migrant children in the US).
Also, where there are several migrants from the same family, the financial burden of sending remittances is shared, which lightens individual responsibility. In the case of parents who have both migrant and non-migrant children, the former are responsible for providing financial and material support, whilst the latter perform the local care of the older generation. Obviously this division of the financial responsibility and emotional caring depends on agreement between the siblings. As Pribilsky (2004) found in another part of highland Ecuador, those families who can convivir (live side-by-side) have better relationships and are more likely to succeed economically and socially. On the whole, Xarbán's remittance dynamics tend not to challenge traditional gender roles but are patterned within them. Men are still cast in the breadwinner role, albeit from afar, and women stay at home taking care of the household and the children. Some reordering, however, does occur in generational relations. Where migrants remit to support their parents, a kind of role reversal occurs. As a father remarked: 'How can I complain [about my children] if we live on them! Before we were the parents, now they are our parents'.
The case of Xarbán should not be generalised too widely, however. Parallel fieldwork by Mata-Codesal (2011) in another Ecuadorian highland community where most emigration is of females to Spain shows much more fundamental gender-role adjustments. Research by Bastia and Busse (2011) on international migration from Bolivia and Peru uncovers a variety of gendered effects. Women's autonomous migration from Bolivia to Argentina shows 'gender gains' to be short-lived. Those who migrate to Spain achieve greater independence but patriarchal relations are reconstituted when return migration takes place. Peruvian migration to the US replicates the Ecuadorian model, being led by men with women left behind to fulfil childbearing, household and caring duties (Julca 2005). Table 5 displays the four main dyads for households in Burim. Unsurprisingly, there is some similarity with the patterns in Xarbán. The crucial difference is that Albanian migration has evolved from a pattern of undocumented males working in Greece in the informal labour market to a subsequent stage where most migrants are legally present (albeit dependent on time-limited sojourn and work permits) and are able to bring in their families. Hence Albanian remittance dyads have been dynamic over time, reflecting the changing structure of the transnational family (Levitt and Lamba-Nieves 2013). Source: Vullnetari and King (2011: 113-123).

Burim and Thessaloniki
Despite this progressive trend towards family-settlement migration, the most common dyad from the household survey remains that from the migrant husband to his village-based wife, who usually has children to look after. Within this main dyadic type there are two variations. In the first, the husband works in Greece all-year-round and makes occasional visits to the village. The relatively short distance involved makes this pattern of visits possible, although the husband's full-time work may limit the time available. The second variant is the seasonal migrant, typically a farm worker in Greece, who alternates up to six months in Greece (the time-limit of an agricultural-work visa) with the rest of the time in the village. According to the survey data, average annual remittances from seasonal migrants are less than half the amounts sent by full-time migrants, a difference which reflects not only the different time periods in migration but also different types of work -the long-term migrant is likely to have a more secure and better-paid job. Below, a village interviewee describes the precarious nature of her husband's employment in Greece: He works in the peach orchards… Then, when that finishes, he goes elsewhere and does welding jobs, whatever he can find, all sorts of work -(Monda, 45, wife of seasonal migrant).
These two subtypes have other differences, more gender-related. Wives whose husbands are away for all or most of the year become de facto heads of household, which is usually seen by them as an extra burden rather than empowerment. For the seasonal-migrant arrangement, the circulating male retains more of the head-of-household role, often integrating seasonal work abroad with farming at home. Another variant occurs when the recipient wife also has to care for other family members -usually her husband's elderly parents. 4 This arrangement reflects the fact that she 'belongs' to her husband's family and has care duties towards them, which in fact may preclude her from joining her husband abroad. If she is a co-resident with her in-laws, then the remittance dyad shifts and becomes from her husband to his father (the third dyad in Table  5) since he is regarded as the household head. However, if the 'patriarch' is too ill (or deceased), the wife and/or mother-in-law receive and manage the remittances. This is the case in the following quote, where the father-in-law, in his 80s, is very sick: My husband brings the money with him when he comes to visit… in April, August and New Year… three times a year… There is no fixed amount, it depends on how his work is going. Myself and my mother-in-law, we women manage it -(Elda, 34, three children and husband's elderly parents to care for).
The second-ranked dyad for Burim (Table 5) is from the son to the father. Where the son is single, this will be the main dyad. Remittances from single young male migrants can be quite substantial, especially if they are bound to their parents by a strong sense of filial duty ; and bearing in mind that they do not have their own nuclear families to support. If the parental household has two or more single sons working abroad, then it is on the receiving end of multiple main dyads, and may accumulate substantial amounts of capital. Regarded as savings rather than remittances, this 'excess' capital will likely be directed towards the selection of a building plot and the construction of dwellings, in readiness for the future marriages of the single sons. Whether these 'remittance houses' will ever be permanently lived in is another question (Dalakoglou 2010).
The fourth dyad in Table 5 concerns remittances from married couples living abroad. Signif icant remittances are not usually expected from migrants living abroad as nuclear families since it is understood that their main financial duties are towards themselves and their children. However, modest sums may be sent to the husband's parents to support them in their old age; such transfers are labelled 'pensions' rather than viewed as remittances. However these transfers are enhanced in their quantity and regularity if the older generation is looking after the migrants' children -an arrangement which allows both migrant parents to work full-time in Greece. Thus far we have described and exemplified the four main Thessaloniki-Burim dyads, all sent by males, either to fathers, or to wives if the migrant is married and has migrated alone. In Burim, as in the rest of Albania, women have become increasingly involved in the migration process, albeit hardly ever as 'independent' migrants. To what extent have they become active as remitters? The answer is simple: to a very limited extent. The traditional Albanian family model excludes this since women are regarded as the 'property' of men (their fathers, then their husbands) and hence income earned by married women in migration likewise passes to their husbands and, if remitted, goes to the husband's parents, not her own. This traditional remittance model was found to be still in operation in an earlier study based on remittances sent by migrants originating from northern Albania (King et al. 2006).
However, interview data from Burim found that women were not entirely excluded from the remitting process (see also Smith 2009). Remittances sent along the female line are nearly always 'secondary' dyads. In fact they are not referred to as remittances at all but as 'presents' or as small amounts of money 'just for a coffee' -phrases which clearly denote their informal and unofficial nature. Usually female-origin remittances are sent or given to female kin, above all mothers and sisters (cf. Rahman and Fee 2009). Female remittances to the wife's parents might also occur if two other circumstances arise: the wife's parents have no sons (such 'daughters-only' families are the subject of pity and sympathy in Albania); or the husband's parents are dead. Irene (37) was living in Thessaloniki with her husband and young son. They used to remit mainly to her husband's parents, but they had died. I don't send them [her parents] money like a pension [i.e. regularly] but whenever some relative goes there [to the village], I would send them €100 or €200 as a dhoro [Greek for 'gift']… Besides money we might buy them clothes, we take them food when we visit… The uses of phrases like 'gifts' and 'just for a coffee' indicate that the conceptualisation and terminology of remittances also needs attention. 5 For now, we return to the main questions underlying our analysis, and interpret them first in relation to the theoretical ideas associated with social remittances and gendered geometries of power; and secondly in terms of the wider debates on the migration-development nexus.

6.
Discussion: social remittances and gender power relations Following Levitt (1998: 933-934), social remittances are 'normative structures [which] are ideas, values and beliefs', and 'systems of practice [which] are the actions shaped by normative structures'. It is easy to conceive of gender relations both as a normative structure (e.g. patriarchy), and as a system of practice -the casual or quotidian encounters and gendered divisions of social roles, including the gendered remittance dyads discussed above. This framework maps onto the two key questions which frame our empirical analysis: firstly the way the normative structures of family and patriarchy shape the initial migration and the follow-on practice of sending remittances; and secondly the way that the transnational gendered social practice of remittance-sending has the potential to reshape gender relations amongst the transnational migrant community, including non-migrant villagers.
Both sending societies -Xarbán/Ecuador and Burim/Albania -are essentially patriarchal and this patriarchality frames the migration process, which was male-dominated from the start in both contexts. Males made up the substantial majority of the initial migrants; this has held true subsequently for Xarbán, less so for Burim where women have joined the early male-only flows. The gendered power geometries of these migrations are very clear; also the inferior 'social location' of women within the local and (through migration) transnational power hierarchies (cf. Mahler and Pessar 2001). Men have orchestrated the decisions as to who should migrate, determining that their wives and daughters should stay at home, or only join them under the gendered and constrained mechanism of family reunion. The latter option has been made possible by regularisation schemes in Greece, but not the US where most Ecuadorians remain undocumented.
Yet, there are indications that patriarchal control over female migration in Ecuador is far from absolute, since there has been extensive female migration from other parts of Ecuador to Spain and Italy (Boccagni 2009;Gratton migration, development, gender and the 'blaCK box' oF remit tanCeS 2007;Mata-Codesal 2011). The conclusion, therefore, is that in Xarbán the male near-monopoly of migration is a combination of patriarchy and the migration control regime of the principal destination country, the United States. For Albania, the patriarchal shaping of migration is stronger. Whilst it is true that in both Greece and Italy women now make up more than 40 per cent of the Albanian migrant populations, nearly all of these are dependent wives and relatives. Independent female migration is still rare in rural Albania. Most young women who move abroad on their own are university students or graduates pursuing higher degrees and professional careers, and come from better-off urban backgrounds (Vullnetari 2007: 44).
Moving to the second key question: what have been the effects of remittances on gender dynamics within the transnational family? For Burim, we observe the following. First, men retain priority in the remittance process, reflecting the overall patriarchy of Albanian society and the more specific patrilinearity of family structures. Second, and countering the first point, some women have been 'allowed', or have insisted, or have resorted to secret means, to send remittances themselves, generally to their parents or sisters. However, the hegemonic male discourse surrounding remittances downplays these female transfers as mere 'gifts' or 'just for a coffee'. Third, the migration of husbands has turned wives into remittance receivers (except where the wife has moved into the household of her husband's parents). If the husband is abroad all-year-round, wives have the responsibility of administering the remittances and the rest of the household -children, livestock, vegetable garden etc. Many complained of these extra economic and management responsibilities. Family separation has emotional costs too, not only for the woman but also for children who lack a father-figure in their everyday lives. The overall effect is to increase women's burden rather than facilitate any 'social relocation' via agency and empowerment.
Many of the points made above also apply to Xarbán. Here too we see a migration process led by men set within a home society where men are accorded more power and status than women. We also note more female decision-making in the management of the household due to incoming remittances and the long-term absence of men in the family. When husbands/fathers return, we find both a wish to reclaim the dominant male role within the family, but also a greater involvement in household chores because of their 'forced training' in these tasks whilst abroad, living in a largely male-only domestic environment. Here is an interview extract from a returnee. Asked whether he has brought back any 'American' habits to Xarbán, he replied: Mmm… well, the habit I brought is that one has to do everything in the house -washing, cooking, ironing [laughs] because… if you want to save money, you have to do all those things yourself. You get used to it. So, here [in Xarbán] I continue to wash-up and cook. This is more than a habit, it's an obligation… and it is much better here because it is in your own home -(Rigoberto, 42, married returnee from Queens, New York).
Another gender-related change in quotidian practice occurs with the migrant household abroad. Below, an interviewee in Thessaloniki describes the reality of household life when the need for both spouses to work brings a reallocation of domestic responsibilities: My wife comes home at six in the evening. Who will cook and clean for me? Who will look after the children? I have to take my share of doing those things… They [referring to people in his home village] don't understand when we explain the tough life in Greece, the conditions we live in… -(Alket, 47, married with family in Thessaloniki).
What is interesting in both these quotes is the extent to which changes in gender norms and practices are the result, on the one hand, of a direct engagement with, and internalisation of, the gender-values of the host society; or, on the other hand, the fact that these changes are forced by the realities of living abroad, either as a functionally single man in New York, or as a hard-working nuclear family in Thessaloniki.

Conclusion
What light do these research results shed on the bigger questions surrounding the links between migration and (under)development outlined at the beginning of the paper? How do remittances contribute to remaking the relationship between poor, remittance-receiving countries and the richer, migrant host countries? Are migration and remittances part of a virtuous or vicious cycle for Albania and Ecuador? The short answer to the last question is 'a bit of both'; but on the whole, we feel that a measure of 'neo-pessimism' (de Haas 2012: 22) is justified.
These issues need to be appraised at two levels: first at the micro-to-meso level of individual migrants and their family and community networks. Both Xarbán and Burim are saturated with the effects of migration: most households have relatives abroad and remittances are the mainstay of the communities' economic existence. For farming families remittances are a complement to food produced locally, perfectly exemplifying the NELM model. For Albania as a whole, an IOM study based on detailed questionnaire surveys in Greece and Italy (N=712) and on household surveys in Albania (N=1066) concluded that, at the household level, 'access to a migration network and family exposure to migration (because of the remittances that follow) is one of the most viable means to escape poverty' (de Zwager et al. 2005: vi). Our data confirm this. However, the downside is that the combination of material and social remittances has mainly worked to generate more migration -or at least to foster an imagination or expectation of migration -amongst the younger generation. Moreover, remittances have reached a stage where they act as a substitute for local economic and employment activity, thereby further depressing the dynamism of the local economy. The main exception is the way that remittances stimulate the local construction sector, reflecting the fervour with which migrants in Burim and Xarbán invest in building new houses in their home villages. Often these houses have symbolic rather than real economic or functional value: they are a physical demonstration of 'success' in emigration and a marker for a future return, although there is no guarantee that such a return, except for holidays, will take place. Although emigration is unlikely to regain the mass-outmigration rates of the 1990s and early-mid 2000s, we foresee that this dependency-cycle of migration, remittances, and more migration will carry forward as long as factors such as limited local employment and career prospects, remoteness, poor infrastructures and unstable local governance continue to bedevil the sending areas.
In terms of the macro scale of the national level and the international political economy of remittances, the dependency relationship is equally clear. The earlier 'dependency-school' characterisation of the Mexico-US migration also holds for Ecuador. Here the various components of neoliberalism -privatisation, deregulation, structural adjustment, and dollarisation -have led to a 'devalorisation' of labour, both in the sending context, where in economic terms it is generated as an unwanted surplus, and in the US where it enters employment niches under conditions of extreme precarity. In this scenario, remittances have a social meaning beyond their monetary manifestation and beyond the Levitt definition of social remittances. Paraphrasing Delgado Wise and Márquez Covarrubias (2011: 65), remittances signify the stretching of the social relations of production across the globe in a context of super-exploitation and social exclusion of migrant workers. Rather than an instrument of development, remittances represent a fraction of wage-labour income designated to cover the subsist-ence of family dependents in the place of origin, including the upbringing and reproduction of future labour migrants.
This neo-Marxist explanatory framework also resonates, to some extent, for the geographically compact Albania-Greece migration and remittance corridor. The dominance of each country in the other's migratory system (most Albanian migrants are in Greece, most immigrants in Greece are Albanians) perhaps implies a more mutual dependency, even a symbiosis (Baldwin-Edwards 2004). Albanian workers have become a structural component of the Greek economy's labour force, socially stigmatised yet willing to do almost any job for low wages. On the other hand, the current contraction and extreme fragility of the Greek economy means that Albanians are the first to feel the pinch as their mostly casual service labour and construction-sector work is dispensed with. In an interesting recent review of the impact of the Greek crisis on Albanian migrants in a small town in northern Greece, Michail (2013) describes a complex picture of adaptation, hunkering down, partial return migration, and above all uncertainty over the future.
Summing up, our paper has offered a rare example of a comparative analysis of gendered migration and remittance dynamics in two contrasting global settings, set within the wide-ranging topical debate over the relationship between migration and development. Deploying the concept of migration dyads and their links to gendered power geometries, we have excavated the main personal patterns of remittance transfers, so often hidden within the black box of monetary sums and spending patterns. Despite differences in geography and family settings, we have found more similarities than differences in the two cases studied.
2. However the introduction of a visa requirement in 2003 left family reunion as the main legal entry route. Like Albanian women in Greece, most Ecuadorian women in Spain work as domestic helpers or carers of elderly locals. 3. The survey recorded information not only on the respondent but also other household members, including those abroad. The 15 per cent figure takes into account the larger average household size in Xarbán than Burim. 4. And hardly ever her own parents, who are the responsibility of her brother(s) and their wives. Traditionally, it is the role of the youngest son to take care of parents in their later years. The youngest son is called 'the son of old age' in Albania. 5. We have dealt with this, based on our Ecuadorian and Albanian material, in another paper (Mata-Codesal et al. 2011: 20-30).