Journey’s End? Old Age in France’s Migrant Worker Hostels

This chapter elaborates the book’s central research question. The current state-of-the art in return migration theory holds that ‘geographically single’ male migrants who are retired or economically inactive will be among the most likely candidates for return to countries of origin in older age. However, the behaviour of the population to which this book is devoted – older North and West African men living in migrant worker hostels in France – is puzzling in this respect. The men’s preference for regular back-and-forth circulation over definitive return to wives and children in places of origin calls into question the current state-of-the-art. Firstly, the hostel residents’ behaviour challenges the assumptions of the ‘myth of return’ literature, which explains non-return on the basis of family localisation. Secondly, the men’s behaviour is puzzling insofar as it is irrational from a neo-classical economics perspective. The men remain unmoved by the financial incentives of a return homewards, where their French state pensions would have far greater purchasing power. The paradox of non-return becomes yet more unexpected when one considers the cramped accommodation, ill-adapted to the needs of older people, in which the men elect to grow old, as shown in the photo essay which concludes the chapter.

Later life is synonymous with reduced mobility in many cultures: gradually older people's physical and social worlds contract as movement becomes more taxing on the body and friends and relatives pass away. A second normative expectation of later life is that younger family members are able to alleviate, to some extent, the physical and social isolation associated with reduced mobility. This has led to the propagation of the widely-held ideal of old age as a time of repose surrounded by one's family.
In this book, I seek to understand and explain a situation which does not conform to this normative image of old age. I am referring here to the experiences of a particular group of older men, of North and West African origin, living in migrant worker hostels in France (foyers de travailleurs migrants). The migrant worker hostels were initiated by the French state in the late 1950s and 1960s. Their original purpose was two-fold: as a means of monitoring a suspect foreign male population at a time of decolonisation and workers' struggles, and as a short-term housing solution for a supposedly temporary migrant labour force. Yet against all expectations the hostels continue to exist today, hosting an ageing cohort of men whose presence in France has proved far from temporary. The men themselves also defy expectations: 'geographically single', they did not bring their wives and children to France while they were working, yet at retirement they do not return definitively. Instead they constantly travel back-and-forth between their hostels in France and their families in places of origin, even in quite advanced states of frailty and ill-health.
As a counterpoint to the image of a sedentary old age surrounded by one's family, the experiences of the older hostel residents are by no means novel or unique. Indeed scholars are increasingly drawing attention to the phenomenon of transnational ageing. King et al. (2000: 31) note how "improved accessibility by both surface and air transport, measured in cost and time" has broadened the opportunities for late-in-life migration. Because of the end of working life and exportability of pensions, retirement marks an appropriate juncture to relocate if so desired. 1 In some countries a clear 'retirement effect' is observable in the return migration rates of older foreigners at or around statutory retirement ages (Klinthäll 2006) and statistics collected by many OECD countries show growing numbers of state pensions paid to recipients who live abroad (Warnes 2009). Mirroring this development, the academic literature on late-in-life migration and transnational ageing has expanded greatly in recent years, including the publication of several monographs (Baldassar et al. 2007;Jovelin and Mezzouj (2015); King et al. 2000;Lulle and King 2016;Oliver 2008), edited collections (Horn and Schweppe 2016;Karl and Torres 2016;Walsh and Näre 2016), and journal special issues (Ciobanu and Hunter 2017;Horn et al. 2013;Jaeger and Madoui 2015;Torres and Lawrence 2012;Warnes and Williams 2006;Warnes et al. 2004).
In a seminal text, Warnes, Friedrich, Kellaher and Torres make an important contribution in highlighting the diversity between categories of older migrants (in terms of wealth, health, legal rights etc.). 2 They identify three main groups of people who migrate later in life: amenity-seeking movers, family-joining movers, and retirement returnees (Warnes et al. 2004). 3 'Amenity-seeking' migration refers to (recently) retired, relatively affluent individuals who move (on a short-or long-term basis) to locations perceived to be conducive to a higher standard of living, thanks to factors such as climate, scenery or cheaper living costs. A sizeable body of literature now exists on amenity migration at retirement, both internally -such as 'sunbelt' migration within the United States (Longino 1992;Rogers 1992) -and internationally, as documented in studies of affluent Northern European retirees who migrate to warmer climes for some or all of the year (Benson 2011;Botterill 2016;Gustafson 2008;Hall and Hardill 2016;King et al. 2000;Oliver 2008;O'Reilly 2000). 'Family-joining' migration refers, as the name suggests, to retirees who move to join adult children who emigrated previously. These seniors are also referred to as 'zero-generation' migrants (Nedelcu 2009) and are often involved in providing informal support to their children and grandchildren (King et al. 2014;Zickgraf 2017). Older people may also migrate when ill-health forces them to move closer to sources of appropriate support, be that formal care in specialist institutions or informal care provided by younger relatives (Baldassar et al. 2007;Rogers 1992). In addition to these established categories, Lulle and King (2016: 3) draw attention to the largely ignored phenomenon of women in the period prior to retirement who "engage in labour migration to improve their lives." The third category identified by Warnes et al. (2004) is return migration at retirement, the focus of this book. Retirement returnees are individuals who emigrated for work earlier in adult life and choose to return to their places of origin following retirement (Cerase 1974). Until recently, this phenomenon was under-researched (Warnes and Williams 2006;Warnes 2009), if not dismissed outright for concerning "insignificant demographic cohorts" (Conway and Potter 2009: 224). The limited quantitative evidence does indeed indicate that definitive return is a minority choice, and that most older labour migrants prefer to 'age in place' (Attias-Donfut et al. 2006;Bolzman et al. 2016). There has thus developed a quite large body of literature on migrants who primarily age in place, with a particular focus on the vulnerabilities which such individuals confront (Buffel 2015;Burholt 2004;Fokkema and Naderi 2013;Ganga 2006;Gardner 2002;Leavey et al. 2004;Samaoli 2007;Torres 2006;Victor et al. 2012;Zontini 2015).
By contrast, in an overview of retirement return migration, Warnes (2009) noted only four publications which focus on ageing migrant populations undertaking definitive return (Byron and Condon 1996;Klinthäll 2006;Malcolm 1996;Rodríguez and Egea 2006). As with other categories, however, recent years have witnessed increasing attention in this area (Barrett and Mosca 2013;Baykara-Krumme 2013;Gualda and Escriva 2014;Sun 2016), including an edited volume on the topic (Percival 2013). Some of this work points to the frustrations and difficulties which returnees encounter in their efforts to re-integrate and adapt to life in places of origin (Barrett and Mosca 2013;Gualda and Escriva 2014;Olsson 2013;Sun 2016). A significant body of literature has also emerged on the determinants of the decision to return, a question of central importance in this book (for an overview see Ciobanu and Ramos 2015). This literature in particular points to the influence of age, gender, property ownership, income, family and social ties, health status, and citizenship acquisition as key influences on return decision-making (Attias-Donfut et al. 2006;Bolzman et al. 2006;de Coulon and Wolff 2006;de Haas and Fokkema 2010;Hunter 2011;Klinthäll 2006;Mesrine and Thave 1999;Razum et al. 2005;Yahirun 2014).
However it is important not to reify the distinction between migrants ageing in place and those who return, as such categories risk obscuring commonalities, limiting the applicability of insights from the wider literature on ageing and migration. Thus Walsh and Näre (2016: 7) propose viewing "return not as a singular migration event but as a structuring narrative marker in transnational lives, through which ideas of home and belonging are negotiated, irrespective of whether physical return has occurred, is projected or is even possible." Furthermore, scholars have drawn an important insight, namely that return should not be conceived as a once-in-time definitive and permanent movement from country B back to country A. Rather, the circular migration strategy of 'bi-residence' -involving regular back-and-forth moves between countries of origin and immigration -is more and more in evidence among ageing migrants (Ahmadi and Tornstam 1996;Attias-Donfut et al. 2006;Baykara-Krumme 2013;Böcker and Balkır 2016;Bolzman et al. 2006;de Haas and Fokkema 2010;Ganga 2006;Hunter 2011;Schaeffer 2001). Warnes alludes to "transnational patterns of residence" by which "older people can exploit, maintain and continue to develop residential opportunities, social networks and welfare entitlements in more than one country," facilitated by low costs of travel and communications (Warnes 2009: 259-360). This can be a 'best of both worlds' strategy, to valorise their "duality of resources and references" (Bolzman et al. 2006(Bolzman et al. : 1361, but may also reflect a situation of 'double absence' (Sayad 1999), "perpetually missing something" (Liversage and Mizrahi Mirdal 2017: 295). Furthermore, the biresidence option can only continue as long as health and finances allow (Lulle and King 2016).
Despite the development of these more nuanced academic categorisations of late-in-life migration, it seems these concepts are yet to be transmitted into public discourse or policy discussions. A more typical approach in lay discussions is that migration and ageing are related in quite different ways. In many OECD nations, there is mounting concern about demographic ageing, perceived as a problem not only for economic growth but also for the increasing welfare burden falling on a smaller working-age population. One solution proposed by some policy makers is for an opening of the borders to younger labour migrants, preferably well-qualified in today's global, knowledge-based economy (Annan 2004;United Nations 2001). A second solution to the growing needs generated by this ageing population is the immigration of younger care workers to look after the native elderly. In sum, when migration and old age are linked, it is usually in terms of younger migrants who are seen as a solution for balancing out the demographic problem of ageing in developed countries. What is less readily recognised is that an earlier generation of labour migrants to Western Europe is ageing (and dying) on European soil, namely those Southern Europeans, Turks, Yugoslavs, and formerly colonised populations from across Africa, South Asia and the Caribbean, who arrived in the three decades after World War 2 (WWII), the epoch referred to in French as les trente glorieuses.

Post-war Migrants to Western Europe: Settling into Old Age
A dominant representation of the post-WWII labour migrations to Western Europe was the youth, if not agelessness, of the migrant workers. "So far as the economy of the metropolitan country is concerned, migrant workers are immortal… they do not age: they do not get tired: they do not die" (Berger and Mohr 1975: 64). This myth of agelessness can no longer be entertained. Those who were once young and gainfully employed have now reached, or are approaching, retirement age. If one had predicted in the 1960s that the ageing of the post-war pioneer generation of migrant labour would take place in Europe, one would have been met with incredulity, so strong were the assumptions -of employers, policymakers, and the migrants themselves -that this presence would be temporary. Yet in time it became apparent that intersecting with the fiction of agelessness of the migrant workers was a second myth, the myth of return (Anwar 1979).
The guestworkers' return dream was put on hold. Inexorably, through a confluence of economics, family dynamics and the unintended consequences of immigration policy, migrant communities began to form in Western Europe (Bolognani 2007;Piore 1979;Sayad 2006). Recession meant that unrealistic savings targets could not be attained, and with the looming uncertainty of continued family separation as stricter immigration controls were introduced, wives and children began to arrive in Europe while they still could (Anwar 1979;Castles et al. 1984;King 1986). The establishment of these new communities provoked a reorientation in motivations -work lost its purely instrumental role, and became a basis for identity and a means of ascribing status within the migrant collective (Piore 1979). Critical in the transition from provisional presence to permanent settlement was the schooling of young children, when "parents find their children speaking German better than Greek or Spanish or Turkish. Then they realise that their children will stay, and that if they themselves return, the family will be irrevocably broken" (Castles et al. 1984).
In countless families, then, the parents' dream 4 of return was pushed further and further back: postponed initially until children had finished their studies; then delayed again until grandparenthood and retirement. Although retirement is a potential juncture for labour migrants to relocate to places of origin, in the case of Europe's guestworkers return at retirement has not occurred en masse. Katy Gardner records how Bengali migrants in London initially envisaged Bangladesh as the spiritual homeland to which they would return at retirement, but with time they began to feel less 'at home' there. Healthcare issues also weigh in their decision to stay in the UK. Furthermore, their children and grandchildren see their future lying in Britain, not Bangladesh (Gardner 2002). Turning to the geographical focus of this book, similar factors are underlined in studies looking at the return decision-making of ageing migrants in France, as I will now elaborate.

Ageing Migrants in France
Survey data highlight the settled nature of the elderly immigrant population in France. Several hypotheses have circulated about the determinants of migrants' decisions to settle or return, but empirical evidence and quan-titative analyses have regrettably remained limited across the main migrant destination countries. 5 As a response to this lack of data, as well as to better know the scale and scope of the needs of older immigrants, in 2003 the national old-age insurance fund (CNAV) 6 and INSEE, the national statistics and economic studies institute, 7 conducted an internationally pioneering quantitative study, the 'Passage to Retirement of Immigrants' (PRI) survey (Attias-Donfut et al. 2006). 8 A significant section of the PRI survey was devoted to the mobility and residential decisions of immigrants at or near retirement.
The survey's principal findings confirmed the settled situation of older immigrants in France. A clear majority of the sample -60% -indicated a firm preference for living out their old age in France surrounded by their children and grandchildren. 25% preferred the circular migration strategy alluded to above, involving regular back-and-forth moves between France and their countries of origin. The study authors referred to this as the va-et-vient option (literally 'coming-andgoing'). Tellingly, only 6% of the sample foresaw a definitive return to their place of birth (Attias-Donfut et al. 2006).
It was argued by the PRI research team that their findings provide strong support for explanations based on the 'implantation' of migrants in a given setting. This theoretical approach which the PRI team integrated directly into their research design was based on the work done by Annie Mesrine and Suzanne Thave (1999) on the return decisions of elderly immigrants. Their aim was to evaluate the 'implantation' of an immigrant in France according to four variables: whether the migrant (i) had a partner/spouse who was also a migrant or was on the contrary French-born (i.e. 'mixed marriage'); (ii) had children resident in France; (iii) was an owner of property in France; or (iv) had acquired French citizenship. Applying these categories to the PRI sample, it was shown that being in a mixed marriage and acquiring French citizenship is strongly correlated with remaining in France past retirement.
The implantation approach bears a strong resemblance to the concept of 'moorings' developed by the eminent American gerontologist Charles Longino: Like boats to a mooring, persons are tied to their environment by investments in their property, by the many community contexts in which they find meaning, by friends and family members whose proximity they value, by the experiences of the past, and by the lifestyles that weave these strands together into a pattern of satisfying activity (Longino 1992: 23).
Longino was writing in the context of older people's migration within the United States, though his insights equally apply to international migration. In Europe, given the importance of family reunification (or family formation) in the establishment of immigrant communities, researchers have put greater emphasis on the location of family members as a determinant of residence decisions at retirement, (Attias-Donfut et al. 2006;Baykara-Krumme 2013;Bolzman et al. 2006;de Haas and Fokkema 2010;Dustmann et al. 1996;Ganga 2006;Liversage and Mizrahi Mirdal 2017;Rodríguez and Egea 2006;Schaeffer 2001;Warnes 2009). For Augustin de Coulon and François-Charles Wolff, location decisions at retirement will depend not only on a comparison of living standards in the countries of origin and immigration, but also on the strength of family relationships (de Coulon and Wolff 2006).
However, my early fieldwork alerted me to the presence in France of immigrants whose migration trajectories do not correspond to the family reunification narrative prioritised by the myth of return literature. I am referring here to those migrants discussed in the French migration literature as faux célibataires or 'geographically single', 9 who have remained apart from their families throughout the long sojourn working abroad. Despite such long absences, these individuals have remained bound to their families through an obligation to send remittances.

France's Migrant Worker Hostels and the Question of Late-in-Life Return
Emblematic of the geographically single situation are the older men of North and West African origin living in migrant worker hostels. On the basis of the four criteria of the implantation scenario outlined above -presence of children in France, marriage with a French-born non-migrant, property ownership in France, and acquisition of French nationality -the hostel residents do not appear to be likely candidates for living out the rest of their days in France: • Being resident in hostels for 'single men', clearly they are not in a married relationship with French-born partners. But this does not mean to say that the men in the hostels are actually bachelors: according to the most recently available census data, close to two-thirds are married to partners who live in countries of origin. Half were already married before emigrating to France (Gallou 2005: 127-8). • Likewise, in terms of property ownership in France, this criterion clearly does not apply to individuals who reside in collective accommodation such as the migrant worker hostels. If they do own property, it is in the country of origin (El Moubaraki and Bitasi Trachet 2006).
• Thirdly, their children have neither been born in France nor grown up there. Yet 72% are fathers to at least one child born in places of origin. One in two residents has fathered four children or more, with the average being 3.8 children per resident (Gallou: 2005: 128). • The fourth criterion, acquisition of French nationality, is a further indicator of their lack of implantation. Based on the 1999 census data, the men in the foyers are much less likely to have acquired French nationality (6%) compared to the migrant population as a whole (33%) (Gallou and Rozenkier 2006: 54).
Beyond their affective ties to countries of origin in terms of marriage, property ownership, children, and nationality, there is an incontrovertible economic rationale for older hostel residents to return home definitively at retirement. I am referring here to the neo-classical economic models by which labour migration has long been analysed. Broadly told, the neo-classical approach has been to analyse migration in terms of the expected costs and benefits to an individual, or put differently, "the response of individuals to economic opportunity at a distance" (Sjaastad 1962: 80; also Harris and Todaro 1970). Wage differentials are the key variable of analysis: in other words, the individual's decision to relocate is based on his or her expectations of wage differentials between countries of origin and destination (Harris and Todaro 1970). This applies to first-time emigrants as well as subsequent migration decisions such as return. Regarding this latter scenario, "for a deliberate return it is necessary that the difference between the benefit and the cost of being in the host country is decreasing over the migration history, and that a point exists where costs overtake benefits" (Dustmann et al. 1996: 226-7).
In a situation of retirement, where employment and earnings are no longer a factor in the decision, the usual neo-classical cost-benefit calculation of wages needs to be reformulated (King et al. 2000). Wage differentials in such a situation are irrelevant. Instead, what is relevant is the purchasing power of one's pension (de Coulon and Wolff 2006;Klinthäll 2006). For example, Byron and Condon (1996) note that in the case of French Antilleans and British West Indians returning to the Caribbean, most of their sample only had the financial means to return following retirement, thanks to the income security inherent in their pensions. Cooperation worldwide between sending and receiving countries on social security means that in many cases pensions are exportable in their entirety, minus any currency exchange fees and deductions made by national social security agencies. For those migrants who have migrated from non-OECD countries to affluent European states, the higher purchasing power of the host country pension in the home country can be taken as a given. Return in such scenarios is the rational choice from the neo-classical perspective.
In summary, therefore, hostel residents appear to be unlikely candidates for living out their days in France, both on affective grounds (family localisation) and economic grounds. And yet, despite their lack of affective ties to France, their retention of ties to the place of origin, and their financial incentives to return, the men in this situation do not return on a definitive basis.
Instead of definitive return, the va-et-vient option is the overwhelming preference of the residents, alternating their time between France and their place of origin. According to the management of 75 hostels surveyed by Françoise Bitatsi Trachet and Mohamed El Moubaraki, 95% of retired residents regularly do back-and-forth trips (Bitatsi Trachet and El Moubaraki 2006: 101). Data provided to me by the largest hostel company, Adoma (formerly Sonacotra), 10 confirms the salience of this phenomenon for older residents, with figures for surveyed hostels revealing that between 80 and 90% of residents over the age of 56 engage in back-and-forth migration. Two distinct mobility patterns at retirement are apparent: the norm for West African hostel residents is that their return visits are less regular but much longer, often 6 months or more. West Africans are obliged to travel by air, and the market dominance of national carriers (and lack of low-cost alternatives) means that fares are considerable. Only the wealthiest West African pensioners are able to afford more than one (long) return trip per year. In contrast, North African elders can choose between competitively priced flights with budget carriers, or the equally economical (but time-consuming) coach-and-ferry combination. These different circular migration patterns are pertinent to the research question insofar as they may be symptomatic of different motivations for the back-and-forth preference, as will now be elaborated.

The Central Research Question
The two theoretical paradoxes mentioned above constitute the basis of the book's central research question. Firstly, hostel residents' decision-making at retirement is puzzling insofar as it calls into question the assumptions of the myth of return literature, which explains non-return on the basis of family localisation. Just like their compatriots who reunified with families in France, the geographically single retirees do not return definitively at retirement. However, the grounds for this non-return cannot be family localisation, since their families remain back home, hundreds if not thousands of miles away. Secondly, their behaviour is puzzling insofar as it is irrational from the standpoint of neoclassical economics. The men remain unmoved by the financial incentives of a return homewards, where their French pensions -paid in euros -would have far greater purchasing power. In the case of North and West African retirees in receipt of a French pension, definitive return is the rational choice since a pension drawn in euros will clearly stretch further in Morocco or Mali that it does in Montreuil or Marseille.
Despite their lack of affective ties to France and their retention of ties to the place of origin, the men in this situation have by and large not returned on a definitive basis. 11 As a result, the 700-or-so hostels 12 operating in France today constitute a predominantly middle-aged, if not elderly, environment. This ageing was already noticeable in the French census of 1990, when the average age of the men in the hostels was 46 years. By the time of the next census in 1999, this figure had risen to 51.7 years (Renaut 2006: 172). Unfortunately, changes to the categories used in the census after 1999 mean that it is no longer possible to distinguish hostel residents from other segments of the population living in collective accommodation such as nursing homes, hospices and longer-term medical care establishments: they are all grouped together under the category of service de moyen ou de long séjour (Croguennec 2012). As a result, up-to-date statistics for France's migrant worker hostels are not available and researchers are obliged to make demographic estimations based on the 1999 census (Croguennec 2012). The data do show however that as of 1999 there was a massive preponderance of long-term residents among the hostel population: 34% of residents first entered this type of accommodation in the 1960s, with a further 36% entering in the 1970s (Gallou 2005). In other words, the hostel population is residentially static and ageing "within the walls" (Renaut 2006: 175). Given the static nature of this population, one can be reasonably confident in estimating the current demographic trends on the basis of the 1990 and 1999 census figures. Two such trends will be noted here. Firstly, an increase in the average age of hostel residents: based on the earlier census data, this is now likely to be in the late 50s or older. Secondly, an increase in the number of residents aged 55 or over: in 1990, 23% of residents were aged 55 or over, equivalent to nearly 22,000 men, yet by 1999 the proportion over 55 years had rised to 51% . Reliable estimates now put this figure between 35,000 and 45,000 (Croguennec 2012;Plard et al. 2015).
The paradox of non-return and ageing in the hostels, already surprising from a theoretical standpoint, is yet more unexpected when one considers some contextual factors regarding the accommodation in which the men elect to grow old. The historical context to the migrant worker hostel policy will be more fully unpacked in Sect. 2.2. Before then I would like to underline three puzzling aspects.
The decision of the French government in 1956 to begin constructing hostel accommodation uniquely for migrant workers, and initially for Algerians only, cannot be disassociated from the contemporaneous struggle for independence in Algeria (1954)(1955)(1956)(1957)(1958)(1959)(1960)(1961)(1962). The hundreds of thousands of Algerians working in metropolitan France at that time 13 were considered to be a security threat by the French Ministry of Interior. Hence one of the core aims of the hostel policy was to keep Algerians apart from the rest of the population, grouping them together in dreary concrete housing at the edge of towns and cities for means of surveillance (Bernardot 1997). A number of observers have noted the 'extra-territoriality' of the hostels, on the fringes of the major conurbations 14 (Desrumaux 2007;Sayad 2006). Frequently, hostels were constructed beside cemeteries, railways, and other undesirable locations, or implanted in the middle of sparsely populated industrial zones (Ginesy-Galano 1984). Fifty years later the hostels tend to remain cut-off from population centres and are poorly served by local amenities and public transport, constituting a form of "spatial discrimination" (Bernardot 1997: 10). The first puzzling aspect therefore is that older hostel residents prefer to remain isolated in these dreary, poorly connected localities (see Fig. 1.1). 13 According to Viet (1998), there were 240,000 Algerians living in France as of 1953. 14 The greater metropolitan areas of Paris, Lyon and Marseille in particular: as of 1999, 53% of hostel residents lived in Ile-de-France (Paris and region), 13% in Provence-Alpes-Côte d'Azur (Marseille and region), and 12% in Rhône-Alpes (Lyon and region). In other words, eight out of ten hostel residents live in one of the three most urbanised regions of France (Gallou 2005). Secondly, a labour market rationale also dictated the location and functioning of the hostels, which were often constructed close to local employers and industrial zones. Once the Algerian War was over, the demands of the labour market began to weigh more heavily than security concerns in the French authorities' approach to migrant workers. The latter were deemed necessary only so long as industrial growth depended on their manpower (Sayad 1999). Marc Bernardot argues that the hostels constituted the housing dimension of a broader labour force policy which aimed "to limit the durable settlement and family reunification of these workers" (Bernardot 1997: 12; see also Ginesy-Galano 1984). On the second count, the policy has been successful insofar as it has inhibited the reunification of many families on French soil. Yet on the first count the residents themselves are evidently well 'settled' in France. If they are not integrated according to customary indicators, it is puzzling that they remain in France for much of the year.
Thirdly, the hostels were explicitly designed as temporary accommodation for temporary workers. As such, the hostels were built quickly and cheaply, according to substandard norms of construction and using materials which were not designed to be durable. The long hours which the men worked meant that the hostel room was usually considered only ever as a place of sleep, needing only to fit a bed and precious little else (Sayad 2006: 94). Hence the miniscule rooms on offer, typically measuring between 4.5 and 7.5 m 2 in surface area (see Figs. 1.2, 1.3, 1.4, 1.5, and 1.6). 15 From being a temporary accommodation solution for young workers, the hostels have become permanent housing (if not 'home') for elderly retirees. The architectural layout and facilities are patently not suitable for the needs of older people who may have cause to spend long periods of time in their rooms: despite this, the men stay on in their small rooms in France.
In sum, then, older hostel residents are not integrated in France according to the conventional reference groups (family, property-ownership, and nationality) and instead retain multiple and durable ties to their homelands. Furthermore, the men remain unmoved by the financial incentives of a return homewards, where their pensions would have far greater purchasing power. Despite these incentives, they do not return definitively, preferring instead to circulate regularly between places of origin and their peripheral, inadequately equipped hostels. These theoretical and empirical puzzles come together in the central research question of this book, namely: 15 The cramped living conditions are improving (slowly) as more foyers pass to the status of résidence sociale. See Chap. 2 for more details of this change.
What explains the hostel residents' preference for back-and-forth mobility over definitive return at retirement? Hostel Living In the following excerpt from one of my interviews, Jawad (69, Taroudant, Morocco) 16 describes his hostel to me. I took the accompanying photographs (Figs. 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9 and 1.10) in the same hostel, inspired by Jawad's words.
Jawad: Ah the hostels, you know the hostels, the people who live there, it's like being in limbo. Like being in a barracks, or down a mine, or like prisoners in jail. It's the same thing.
Jawad: Well the room, you know the room. AH: Yes of course. Jawad: Just 7 square metres, AH: Not even! 16 Henceforth, the age and origin of all hostel residents cited in the text is formulated according to the following character key: pseudonym; age; region of origin; country of origin.  Jawad: … also, everyone has a small locker where they can put their groceries, potatoes, onions, pans, cups etc. Fig. 1.6 The shared kitchen, with gas burners and sink If you arrive late, and find all the stoves taken, you have to wait … Jawad: … one hour, two hours, it depends -because some people don't eat until half ten in the evening. AH: At least you have a nice view to look at while you're waiting. The toilets -communal. The kitchen -communal. Even the bedcovers and sheets -communal! You don't have the right to anything!

Purpose, Title and Structure of the Book
As of 1999, when the last census measuring the hostel population was conducted, the hostels housed almost 80,000 male migrants, amounting to only 3.6% of all foreign-born men in France (Gallou 2005). Given this numerical insignificance, and their particular family situation (i.e., geographically single), the hostel residents are far from being representative of all older foreign-born individuals in France (Gallou 2005). Tony Warnes, who has contributed significantly to the study of older migrants, writes that "[t]he attention of journalists and researchers is drawn to the more unusual forms [of migration], which are visible and available for study but unrepresentative" (Warnes 2009: 343). This critique undeniably applies to the population selected for study here. In light of this, readers would have legitimate grounds for asking: 'Why is this worthy of our attention? What contribution can this research make to the migration studies field?' To respond firstly by way of intellectual justification, my interest in the hostels initially came about during earlier fieldwork in France, through interviews I was conducting with migrants' associations in Paris. In response to my question about what the priorities for their associations were, several representatives highlighted the topic of the older men living in migrant worker hostels, and their problematic health and welfare situations. Intrigued, I subsequently discovered that there was very little known about these institutions outside France. While drafting this book I was able to source only three English-language references devoted primarily to the hostels (Jones 1989;Diop and Michalak 1996;Mbodj-Pouye 2016). More intriguing still was my discovery that much of the major francophone scholarship on hostels has been conducted by former employees of the main hostel companies. 17 From the outside, it appeared to me as if the hostels constituted a 'dirty little secret', knowledge of which has been controlled by certain vested interests. 18 Furthermore, other literature which in places alludes to the hostels is misleading. In their widely cited Here for Good: Western Europe's New Ethnic Minorities, Stephen Castles and colleagues argue that in terms of housing provision in Western Europe, "[a]part from a few restricted schemes in France and West Germany, very little in the way of housing or social amenities was provided especially for migrants" (Castles et al. 1984: 29). Under any circumstances, it is hard to concede that a total complement of over a quarter of million hostel beds for migrant workers constitutes a "restricted scheme," as was the case in the mid-1970s heyday of hostel construction (Lévy-Vroelant 2007). 19 In addition, earlier literature on return migration tended to focus on the impact of emigration and return on places of origin (King 1986;Cerase 1974). This tendency has continued in contemporary analyses of what is known as the migration-development nexus. Against this trend, Laura Jeffery and Jude Murison have drawn attention to a less discussed issue: the impacts of the return process on the returnees themselves. As they note, "[r]eturn and re-integration into the home 17 Jacques Barou, Marc Bernardot, Mireille Ginesy-Galano and Choukri Hmed have written extensively on the hostels and prior to becoming academics worked for Sonacotra-Adoma, the largest hostel company. The latter three are responsible for the major reference works on the topic. I do not mean to suggest that these scholars' work has been compromised by their prior association with Sonacotra-Adoma: indeed, in some cases it appears that this scholarship is a reaction against the experience of working for the company. What I would underline however is the idea that insider contacts seem to have been necessary in the past to do academic research on and in the hostels. 18 This impression was only reinforced when I attempted to negotiate permission to undertake research in hostels managed by Sonacotra-Adoma. My overtures to the company's head office met with delaying tactics and, ultimately, refusal to respond. Access was eventually secured at a much more devolved, local level. 19 Lévy-Vroelant (2007: 21) enumerates 264,800 hostel beds as of 1975. Given the fluid definition of what counted as a migrant worker hostel, other estimates are somewhat lower. Patrick Simon (1998: 46) enumerates 680 hostels operating as of 1974, housing some 170,000 workers. country may be rife with difficulties for returnees" (Jeffery and Murison 2011: 132). While this book does not neglect the former concern with returnees' impact on their places of origin (see Chap. 6 particularly), above all I am concerned with hostel residents' perspectives on their own return decisions.
Empirical gaps aside, this book also aspires to make a substantive theoretical contribution. Arguably the lion's share of the francophone literature on the hostels has focused -not without eminent justification -on hostel residents' alarming health and welfare situation (Boyer 2001;El Moubaraki and Bitatsi-Trachet 2006;Gallou and Rozenkier 2006;Samaoli 2007). While my focus on return does not prevent me from considering such matters (see Chaps. 4 and 7), I would argue that the hostel residents also constitute a population which can generate significant new insights for migration theory, despite their numerical insignificance. Indeed, the theoretical value of this research lies in the very fact that it takes an anomalous case. Of all the post-WWII labour migrants to France, the hostel residents would appear at first sight to be the most likely candidates for return. The myth of return literature has relied on family localisation as the principal explanation for the lack of return amongst the post-WWII immigration cohort in Western Europe, yet this explanation cannot apply here. Just as fundamentally, the decision not to return definitively calls into question neo-classical economic theory.
What then can account for this paradox? In Chap. 2 I present four distinct theoretical frameworks which may provide answers and help us to understand and explain late-in-life mobility. The new economics of labour migration proposes that return decisions are part of wider household strategies to minimise risks to family income (Stark 1991). Structuralist accounts (e.g. Cerase 1974;Sun 2016) draw attention to the political, social and cultural structures in places of origin to which returnees have to re-adapt. The literature on transnationalism (e.g. Portes et al. 1999) by contrast stresses the ties which migrants retain both to their host countries and to their homelands, and the evolutions in technology, transportation and governance which facilitate such ties. Finally, the theory of functionally differentiated social systems developed by Niklas Luhmann (1990) gives priority to the role of the welfare state in mediating the exclusionary tendencies of modern society. Access to healthcare and other forms of social protection becomes more important in later life and may have a bearing on return decisions.

Title and Structure of the Book
The question of continued residence in hostels past retirement is what I seek to explain in this work. The original title of the research project from which this book derives was: 'Retirement Home? France's migrant worker hostels and the puzzle of late-in-life return.' As I have tried to convey in this Introduction, there was something initially very puzzling about hostel residents' non-definitive return at retirement, from the point of view of the family localisation thesis as well as the argument from pension purchasing power. However, in time I came to understand various aspects of the research question better. Instead of a puzzle, I began to appreciate the issue from the perspective of the residents themselves, as a dilemma, given the difficult choices they have to make between unfavourable alternatives. Reducing their experiences to the level of a theoretical 'puzzle' would risk dehumanising and trivialising what can be a highly fraught situation.
From time to time the human drama of the hostel residents' lives is the subject of public discourse and controversy in France. Headlines like "Immigration: the forgotten hostels" (Decugis 1998; author's translation); 20 "The refuge of the uprooted" (Guyotat 2004; author's translation); and "Forgotten, more and more immigrants are growing old alone in France" (Bernard 1999;author's translation) give an indication of the tenor of this discourse. Other articles note with alarm that the Sonacotra hostels are "transforming into retirement homes" (Serafini 1999; author's translation), or worse, mouroirs -a pejorative term implying a refuge of last resort, where one goes to die (Guyotat 2004;Decugis 1998).
Such headlines are the inspiration in part for the interrogation found in the title to this book, Retirement Home? This interrogation works on several levels. Firstly, it speaks to the question of whether the hostels constitute an appropriate living space for older people: can we really consider the hostel to be a retirement home like any other? Secondly, it speaks to the difficult question the men face at retirement: to return home definitively or not? Thirdly, in this book I question what retirement means for these ageing migrants and find that in several ways their passage to retirement does not correspond to normative expectations of later life, be it in France or in places of origin. Finally, I conclude by asking what 'home' can mean for the hostel residents: are they 'homeless' everywhere, as the more pessimisticallyminded scholars conclude (e.g. Sayad 1999), or have they found home in new and unexpected places?
Chapter 2 sets the scene, reviewing the history of the hostels and the relevant theoretical literature, before providing details of my multi-sited research design spanning France, Morocco and Senegal. At this stage I underline the importance of the concept of biography for this work. In Chap. 3 and thereafter, key elements of residents' biographies are characterised as deviating from the 'standard' lifecourse expected by various welfare state institutions. Now that they are older and no longer working, the men's biographies -recorded in documents such as passports, social security files and payslips -mediate their difficult relationship with the French state, notably in terms of social security benefits and healthcare.
Accessing these two goods is a principal rationale for the retired hostel residents' preference for the va-et-vient over definitive return, as is elaborated in Chaps. 3 and 4 respectively. Many social security benefits are subject to minimum residence conditions. Administrators at various state agencies seek to territorially 'fix' and temporally 'timetable' the hostel residents through strategies such as passport checks, tax declarations, and targeted fraud investigations. With healthcare, the betterquality, subsidised services available in France mean that it is the preferred location for most treatments. This implies that a large burden of care falls on the French welfare state, since the men have no family in France to look after them.
Instead, hostel residents' families are transnational. In Chap. 5, it is shown how fundamental these family ties are for respondents, as manifested in their remittance sending practices: it was the dream of family prosperity and financial security which spurred their emigration in the first place. Unexpectedly, the families of some hostel residents remain dependent on remittances even following the emigrant's retirement, leading to prolonged stays in France in order to claim certain social security benefits which are subject to a minimum period of residence being observed. However, the long period of exile can lead to loss of influence within the family, especially for North African respondents. From the perspective of hostel residents, this undermines patriarchal norms in gender roles and family relationships.
Chapter 6 discusses the possibility of re-integrating to communities of origin. Many North Africans complain that they have 'lost their bearings' in their families and in the wider home community, thus rendering difficult their prospects for reintegration there. Most West African men, it transpires, are better able to re-integrate thanks to their involvement in hometown associations. These structures enable them to maintain a political, social, and economic presence despite their long absences, facilitating re-integration when the men come to retire.
A further function of hometown associations is to provide repatriation insurance in case of death in France. Chapter 7 documents how hostel residents approach physical frailty and the end of life. The va-et-vient can continue only so long as the men are in a fit state to travel: eventually, ill-health and/or loss of autonomy force the men to choose where to live out their days. The options in France are unappealing. In terms of facilities and architectural layout, most hostels are an entirely unsuitable environment for older people with advanced dependency. Yet to enter a dedicated residential care home is to renounce the remittance sending role, given that the fees charged are so high. Hence some undertake a 'penultimate voyage' homewards -returning home to die -in order to benefit from family care in their last days and also to be assured of a funeral in accordance with Islamic rites. 21 That such rites might not be properly followed in France is a source of existential anguish to many.
The book concludes by offering an answer to the guiding question of why the hostel residents prefer back-and-forth mobility over definitive return at retirement. Given the limited explanatory value of the arguments from neoclassical economics and family localisation, I assess the merits of the rival theories which were proposed: the new economics of labour migration, structuralism, transnationalism and social systems theory. In the light of the findings of Chaps. 3, 4, 5, 6, and 7, no one theory adequately accounts for all the phenomena observed: at various points in the data there is support for different theories. Nonetheless, I argue that Luhmann's theory of social systems has the greatest potential for fruitful application in the future, since it offers a radically fresh perspective on why people migrate and the structure of the society into which they integrate. This insight is developed by applying an innovative systems theoretic approach to an idea which features prominently in migration research, including in the title to this book, namely the concept of 'home'. Building on conventional conceptions of home predicated on bonds to social group or territory, I argue that to be 'at home' can also mean upholding claims to be 'included' in different social systems. This argument is particularly salient in light of a recent policy measure offering a guaranteed monthly income to hostels residents who return to countries of origin, and in the book's final section I consider the prospects for this legislation and the ageing residents which it targets.