The Making of Cheap Labour across Production and Reproduction: Control and Resistance in the Senegalese Horticultural Value Chain

Abstract

This article investigates the making of cheap workers at the bottom of global value chains. Adopting a class relational approach, it engages in labour regime and social reproduction analyses, to examine the labour process in Senegalese export horticulture and its relations with rural households. Drawing from primary qualitative data, it analyses power relations within workplaces and households investigating the structural relations between them through the disciplining and exploitation of women. It argues that labour control beyond workplaces is crucial to the supply of cheap and disciplined workers, showing how patriarchy and religion regulate a continuum of class relations between households, fields, and packaging centres. It shows the inherent conflict between production and reproduction intensifies South of the supply chain and fuels fragmentation of women in ‘classes of labour’. While women shape, respond and defy some forms of subordination, their resistance to the combined pressures of disciplining and exploitation is less manifest.

Introduction

Since the 1980s a broad literature has documented a widespread, though uneven, ‘feminisation of labour’ underpinning the relocation and expansion of manufacturing and food production across the global South (Elson and Pearson, 1981, 2011; Jackson and Pearson, 1998; Mies, 2014) based on the seemingly boundless availability of cheap female labour, pervasive in export agriculture (Deere, 2005). Here, ‘far from searching for “cheap” labour, as today’s globalisation-friendly story tells us, structural adjustment policies and other international trade agreements effectively commanded developing nation-states to make ever cheaper labour present (and present labour cheaper) in exchange for loans, capital, military aid and a place in the “family of nations”’ (Staples, 2013: 21, emphasis in original). This article contributes to a closer understanding of how women working at the bottom of global value chains continue to be made cheap today. It explores ‘how gender, as a set of context-specific meanings and practices, intersects the structure of global capitalism and its systemic logic of value extraction and capital accumulation’ (Bair, 2010: 205). It goes beyond the analysis of the labour process to examine the broader labour regimes structuring this cheap labour supply. Using recent fieldwork data from Senegal, the article extends the boundaries of the value chain to include rural households and the unpaid work therein, and follows women between their home and workplace tracing multiple forms of control and some of its responses.

The article responds to calls, made in this journal, for a better understanding of the nature of work and society in the global South and its relations with the global North (Beck et al., 2016). According to Mies ‘we cannot close our eyes to the stark fact that women of all classes in the West, and middle-class women in the Third World, are also among those whose standard of living is based on the ongoing exploitation of poor women and men in the underdeveloped regions and classes’ (2014: 1). Thus, the article asks how does this exploitation work in specific places? What does it depend upon? And whether, and how, is it challenged?

To answer these questions the analysis builds on the emerging dialogue between labour process theory (LPT), global value chain (GVC) and global production network (GPN) studies (Newsome et al., 2015) that has prompted new interest in labour regimes. The article combines labour regime and social reproduction analyses to highlight the constitutive connection between capitalist production and social reproduction and recast the central role of households for a deeper understanding of exploitation. Whilst one of the earliest definitions of global commodity chains puts households at its heart (Gereffi et al., 1994), the early stream of GVC literature overwhelmingly focused on governance (Bair, 2005; Bernstein and Campling, 2006; Smith et al., 2002) and later moved to production and labour though largely within the walls of factories/plantations (Barrientos et al., 2011; Dunaway, 2014; Kelly, 2009). Most recently, despite a ‘social turn’ (Reinecke et al., 2018) and expanding/deepening GVC/GPN analysis (Coe and Yeung, 2019), household dimensions remain largely unexplored. Yet, household analysis can contribute to these scholarships in several ways.

A focus on households highlights who does what job, under what conditions and trajectories – and why – thereby emphasising how the making of workers transcends the factory, the office or the field: exploitation is simultaneously the generative and culminating act in the relation between capital and labour that builds on a hidden forest of relations crossing workplaces, households and other places/institutions. By trying to unbundle some of these, the article challenges much of the workplace and network ontologies inherent to LPT and GVC analyses (Newsome et al., 2015). Indeed, household analysis also provides a terrain to investigate the hidden, indirect, role of the state in reproducing a segmented, class-based society beyond states’ explicit policies identified by GVC literature (Lombardozzi, 2020). Indeed, both scholarships can be enriched by drawing from feminist agrarian political economy that takes the household as a point of departure rather than an add-on (Razavi, 2009). This perspective also contributes to feminist critiques of LPT based on a relational understanding of the division of labour across spheres and places (Glucksmann, 1990, 1995, 2005) as here the ‘relational’ is first and foremost a class relational approach (Campling et al., 2016) where women occupy different positions within vastly fragmented ‘classes of labour’ (Bernstein, 2010). Compared to downstream households, the rural household South of the GVC occupies a different position in the circuit of capital as net supplier of labour for global production (whether directly in factories and plantations or indirectly through cash crops). Its women are not substantial consumers of reproductive commodities (whether a washing machine or a carer’s labour) and welfare: their intense exploitation prevents them externalising the market key reproduction costs, neither does the adjusted-state fill the gap (Ossome, 2015). Rather, in the South the inherent conflict between production and reproduction is heightened (Fraser, 2016; Naidu and Ossome, 2016), and this article contributes to an understanding of how this is managed (or unsolved) in Senegal.

Senegal, a major West-African supplier of off-season fresh fruit and vegetables to Europe, shows a widespread and enduring feminisation of the workforce sustaining the development of horticultural value chains since the 1970s (Mackintosh, 1989). Drawing from qualitative data collected from a sample of women working in packaging houses, the analysis unveils a labour regime based on the disciplining and exploitation of women, at home and in factories, that reflects the development of capitalism in the countryside and its crucial mediation by households as the echo chambers of other institutions, including the state, religion, and polygyny. Here labour control outstrips the labour process, operating across firms and households and structuring the reciprocal link between work at home and work in factories. It reveals that large households and informal, seasonal, poorly paid wage employment reinforce one another. It examines some of the most manifest forms of (apparent and actual) resistance, inside and outside firms, and the implications for a broader understanding of workers, their fragmentation, and a relational understanding of class struggle. This illustrates the multiple facets of the latter as women juggle both ‘struggles as women’ and ‘struggles as workers’ (Elson and Pearson, 2011), a reflection of class operating through gender as one of its fundamental ontologies (Federici, 2004).

Following this introduction, labour control regime analysis is adopted to reconnect theoretically capitalist production and social reproduction framing labour control along the interplay between exploitation and disciplining. Then the construction of gender in rural Senegal is examined, and the case study and data collection process are introduced. The next two sections analyse forms of control and resistance within households and factories. Some overall arguments are advanced in the conclusions.

Integrating labour control and social reproduction analyses

A crucial insight of labour control regime analysis is that capitalist production relations outstrip the labour process. This was a core argument of Burawoy’s (1985) notion of ‘factory regimes’, where the labour process was one element within broader production architectures alongside the joint interaction of market forces, capital in competition, capital labour relations, and significant state intervention. Labour control regime analysis further developed within LPT (Thompson and van den Broek, 2010), but without systematically including the sphere of social reproduction. An early step in this direction came from within geography and Jonas’ notion of local labour control regimes as emerging from ‘the time-space reciprocities between production, work, consumption and labour reproduction within a local labour market’ (1996: 325). Despite this intuition, a full integration of social reproduction into labour control regimes analysis remains marginal (Baglioni and Mezzadri, 2020).

Within the growing dialogue among industrial relations, development studies, and economic geography, a second generation of studies furthered labour control regime analysis within the theatre of global supply chains and from the perspective of capital-state dynamics (Anner, 2015; Pun and Smith, 2007), their spatiality (Baglioni, 2018; Kelly, 2002; Smith et al., 2018), and dimensions (Pattenden, 2016; Taylor and Rioux, 2018). A few emphasised the constitutive relations between production and reproduction (Baglioni et al., 2018; Helms and Cumbers, 2006; Mezzadri, 2016; Neethi, 2012; Pattenden, 2020; Pun and Smith, 2007) and documented capital’s ability to systematically externalise the cost of social reproduction to gendered and racialised workers (Dunaway, 2014; Mezzadri, 2016; O’Laughlin, 2013). Indeed, as feminist political economy has shown, far from a discrete realm regarding workers’ free time, the reproduction of workers is a crucial component of capital accumulation and surplus value extraction.

Social reproduction analysis recognises production and social reproduction as a unitary process in capitalism (O’Laughlin, 1977) and de-fetishizes the realm of social reproduction as natural to analyse it instead as a capitalist process, which re/produces labour power. Historically, the making of the working class relied on the gendered division of labour between productive and reproductive activities in Europe (Federici, 2004) and enslavement and housewifisation in the colonies (Mies, 2014). Thus, the making of the working class operated through the construction of difference (Elson, 1999) to subsume labour in different ways and across different positions within the circuit of capital: labour internal differentiation (across gender, race, ethnicity, religion, nationality, etc.) is constitutive of ‘classes of labour’ (Bernstein, 2010).

One of the crucial instruments to the making of the global, gendered, working class, was and continues to be the home (Kuhn, 1986). Largely alienated from production and relegated to the ‘private sphere’, household relations have rarely been considered a constitutive element of the labour process, of global value chains (Dunaway, 2014; Kelly, 2009), or of traditional Marxist analysis where ‘everything seems to happen as if men experienced their alienation and their reification first in their own work, whereas in reality each one lives it first, as a child, in his parents’ work’ (Sartre, 1968: 62, emphasis in original). Yet, by mediating the universal class and the individual (Sartre, 1968), and catering for the daily and generational reproduction of workers, the household stands in a crucial relationship with the labour process and capital accumulation. Moreover, because ‘social reproduction’ straddles the reproduction of workforces and of class relations (Bhattacharya, 2017), the home is simultaneously a core site for reproducing and disciplining workers, a space that historically ‘concentrated the effects and targeted efforts of the state, philanthropists, and social elites to structure and fashion order among the labouring classes’ (Staples, 2013: 55). Thus, the home becomes a locus of spatial control, which ‘whether enforced through the power of convention or symbolism, or through the straightforward threat of violence, can be a fundamental element in the constitution of gender in its (highly varied) forms’ (Massey, 1994: 180). Hence, the need to see households as part and parcel of labour regimes, where class, gender and race are enforced and fought over (Davis, 1983) as in the workplace.

To combine labour control and social reproduction analyses and structure the continuity between household and work relations, I frame labour control along the interplay between labour exploitation and disciplining (Baglioni, 2018). Whilst labour exploitation corresponds to capital extraction of value from the labour process, labour disciplining includes all those mechanisms that mitigate, contain and prevent the conflict inherent to production and is therefore a crucial ingredient in the making and remaking of workers – whether they lie at the centre or the margins of the capital circuit, in factories, offices, fields or trucks, and at home. Just as labour exploitation serves the production of commodities, labour disciplining serves labour exploitation: the two are necessarily linked and often mutually reinforcing. Moving from the idea that ‘capitalist discipline operates through a variety of control mechanisms in social, political, and work domains both to regulate and legitimate unequal relations which sustain the process of industrial modernisation’ (Ong, 2010: 4), labour disciplining refers to the constant direct and indirect subordination of labour to the capitalist labour process and is therefore far more wide-ranging than workplace disciplinary regimes (Edwards, 2010; Edwards and Whitston, 1989). It is indeed a form of class struggle from above (Hanlon, 2016), and the construction of gender as a crucial dimension within it.

Inherently dynamic and contingent, disciplining is manifest in social, economic cultural, spatial and temporal dimensions and always bears the footprint of major institutions, first and foremost the state. The combined nature of disciplining emphasises that exploitation requires multifaceted workers subsumption – i.e. a whole constellation of ways and forms that workers are directly or indirectly made dependent on producing value for capital. The deconstruction of how workers are disciplined in different ways and within different spaces/times reconciles concrete (more or less manifest) forms of oppression, coercion and marginalisation to abstract capitalist relations of production allowing to unravel that forest of relations through which workers become in different ways ‘capital-posing labour’ (Banaji, 2011). Thus, the analysis of labour disciplining requires investigating how gender, as one socially constructed source of difference, is inherently contingent and mediated (by the state and other institutions) and producing distinctive ideas of man/womanhood in different places, and different ways these are challenged (Hart, 1997). While labour disciplining already expresses the contested nature of control springing from ‘the dialectical relation of capital and labour’ (Edwards and Whitston, 1989: 7), not all agency effectively challenges disciplining and exploitation. In other words, while workers always shape the capital-labour relation, they are far less able to resist it. Thus, as women shape, challenge and rework different forms of disciplining encroaching them, like all workers everywhere they are less able to consciously ‘resist, subvert, or disrupt’ the conditions of exploitation and oppression inherent to capitalist production (Katz, 2004: 251). This highlights the multiple sources of disciplining facing women – and crucially the contradictory requirements of social reproduction and production (Fraser, 2016) – as challenge to one might imply greater reliance to the other.

Before focussing on labour control and resistance in Senegalese households and factories, the next section sketches some core elements of the construction of gender in rural Senegal. ¹ These processes combine pre-colonial, colonial and post-colonial legacies, including the institutionalisation of polygyny within religion and the state, and – as argued below – its mobilisation to advance capitalist relations. In other words, the disciplining and exploitation of women are mediated by their agrarian context, the state, and religion.

The construction of gender in rural Senegal

In sub-Saharan Africa ‘households’ are generally portrayed as a complex set of relationships and divisions of labour that challenge prevailing distinctions between productive and reproductive activities (Crehan, 1992; Dancer and Tsikata, 2015). While the literature on intrahousehold relations documents a vastly rich range of situations, there is general agreement that households are characterised by conflicts in the division of labour and resources (O’Laughlin, 2007; Whitehead, 1981; Whitehead and Kabeer, 2001). The complexity of household relations somehow reflects a long-lasting historical trajectory that put them at the heart of the colonial and postcolonial states’ effort in reproducing the conditions for continued capitalist development and accumulation (Carney and Watts, 1991). While successive waves of capitalist development have affected household relations and their sexual division of labour in different ways, the underlying picture is one of intensification of women’s social reproductive work as both the colonial and postcolonial states were ‘complicit in efforts to sustain patriarchal structures of power which underlay the local peasant economy’ (Mbilinyi, 2016: 118).

While Senegal’s ‘peasant economy’ reflects this broader history based on women’s subordination as quintessential ‘reproducers’, one of its distinctive features is the resilience of polygyny (Oheneba-Sakyi and Takyi, 2006; Therborn, 2006). ² In Senegal polygyny is still widespread across different classes (Antoine and Nanitelamio, 1995; Gning and Antoine, 2015; Kringelbach, 2016); it was regularised and legitimised by Islam (Creevey, 1996) and linked to the historical relation between Islamic brotherhoods, slave trade, colonialism and capitalist development (Barry, 1988; Copans, 1988). A complex, flexible and diversified institution, ³ polygyny still represents a ‘solution’ to the crisis of social reproduction where most of the burden is shouldered by women. Overall in Senegal, the social status of women is primarily defined by their marital status and their ability to have (more) children (Bass and Sow, 2007). Amongst the Wolof (the largest ethnic group in the country and in the sample) the idea of woman is deeply entangled with motherhood, care, and the home, her archetypal virtues are patience and obedience to husbands/parents (Diop, 2012; Mbow, 1997, 2001).

In Senegal therefore, the development of capitalist relations was able to include, rather than dissipate, polygyny, reflecting how ‘historically, as wage work has spread, capital has seized upon pre-existing division between men and women, and has incorporated that division within its own workforce to its own advantage’ (Mackintosh, 1981: 6). Underpinning this incorporation is perhaps the particular coherence between its specific gender relations and the capital-labour relation. More specifically, ‘the institution of polygyny determines a particular type of social relations between sexes where particular male techniques and tactics are exercised. These play on the use of threat, secrecy, and the preservation of competition among real or potential “rivals”. The “blackmail of polygyny” is a way of saying to the woman that she is replaceable’ (Fainzang and Journet-Diallo, 1991: 220). Competition amongst women as wives/mothers is always present, even in monogamous unions where ‘women are subjected to the latent risk of becoming [polygynous]’ (Antoine and Nanitelamio, 1995: 20). In this context, household relations highlight women’s, real or possible, replaceability in the home as a powerful pressure to intensify their performance and work. The parallel with labour replaceability within low skilled jobs is striking and not casual. Indeed, in Senegal this strong regime of female disciplining based on the construction of their role, rivalry and replaceability is indirectly endorsed by the state.

By legally recognising polygyny, as the man’s right to marry up to four women simultaneously, the Senegalese family code ‘validates the patriarchy already present in African society’ de facto institutionalising female subjugation (Bass and Sow, 2007: 87). Thus, women are second-class citizens (Creevey, 1991) of a state that is inherently masculine because it secures ‘the right to equality among men’ and profits from female ‘work in looking after the household that is qualified as natural, hence cost-free’ (Sow, 1997: 3). Indeed, neither the state, nor democracy are gender-neutral (Razavi and Jenichen, 2010). The same applies to ‘development’: the promotion of austerity measures and small-scale farming by international and national institutions magnifies the burden of care for women and continues to blur home/workplace boundaries, de facto increasing female work endlessly inside and outside domestic walls (Diaw, 2004). This limitless demand for female labour – where care, reproduction and production structurally overlap – is inscribed in the development of Senegalese export horticulture since the 1960s. As shown below, while not exempt from contradictions, the large-scale employment of women in the sector does not fundamentally challenge their reproductive role, but instead fits in with their status as wives/mothers. Export horticultural jobs join a plethora of ‘income generation activities’ that women pursue as a supplement to and break from their household activities (Mies, 2014).

Research context and methods

A pioneer of African export horticulture, Senegalese exports have relied on female labour since the 1960s (Mackintosh, 1989). The sector has evolved through different phases (Baglioni, 2015), and since 2010 a new wave of foreign capital is expanding its frontier, both North and East of the Niayes area, historically the heart of export horticulture. Despite its evolutions, women remain key to the sector and represent the bulk of the workforce. The vast majority fills unskilled employment that peaks at harvest (December–May) for picking and packaging, the former in fields and the latter, the focus of this article, in factories.

The analysis relies on data collected in two Niayes villages (East of Dakar) during the 2017–2018 agricultural season, following research started in 2007 (Baglioni, 2009, 2015, 2018). ⁴ First a survey was undertaken with a random selection of 90 women employed by export farms. The survey identified women’s age, origin, and marital status, households’ size, composition, economic activities, and sources of wealth, ⁵ as well as basic information about their employment (e.g. current and previous employers, length of employment, tasks, payments). Second, the data collected served to select a representative smaller sample of 20 women employed by larger exporters supplying European supermarkets. Women were selected on the basis of different ages, marital status, household size and wealth. The selection reflected both the broad differentiation among women in terms of age (from 15 to 64), marital status (unmarried, first/second/third wife, widow) and household life-cycle, ethnic origin (largely Wolof), women schooling, and ‘household economy’. The latter was largely based on combining horticultural production, male employment predominantly in transport and constructions, and female employment in horticultural factories and petty trade and services. Life histories and semi-structured interviews were conducted with the sample to explore household internal relations and differentiation. Interviews with these women focussed, on the one hand, on reconstructing the core stages and trajectories of their lives, their daily routines, the division of labour within the household, and, on the other hand, on factory work and forms of control, their reactions to problems and collective organisation within and beyond work.

The women selected for interviews predominantly worked at two of the country’s earliest and largest export horticulture farms, which at harvest employ thousands of women across fields and factories. In one village women were employed by a more distant farm using company buses, whereas in the other village women worked at the factory headquartered at its centre. Reflecting the broader employment structure of the sector, almost all women interviewed were daily workers paid by the piece, a couple were paid daily and one was a supervisor. All interviewees worked seasonally in the packaging of French beans (December-May), only a few were employed for mango processing (June–September). As discussed next, their employment in factories cannot be fully understood in isolation from their position/role in the family and the forms of control therein.

Control and resistance in the household

Most of the women interviewed lived in large, complex and rather fluid households characterised by intricate power relations and hierarchies. Households in the sample varied from 5 to 30 members with an average of 17 members living in the same compound, the size often reflecting the life-cycle of the households, the smaller generally representing newly generated families. All households represented internally complex hierarchical worlds segmented by gender, age, religion, available resources and revolved around the figure of the (male) head.

The sexual division of labour was fairly consistent: men were employed full time and/or were petty commodity producers, whilst women would be responsible for all housework, work in the household fields if available, and take paid work if all the above was fulfilled. While the household heads were generally responsible for the economic support of the entire household, women could be individually responsible for their own children. Overall, the burden of domestic work was female: women were therefore disciplined to be first and foremost housewives, itself a fairly complex category as described below.

Among women, relations were complex and organised according to proximity to the household head, marriage relations, age, and cyclical power structures (Kandiyoti, 1998), where older women exerted considerable authority over younger ones. Domestic work – a diverse series of tasks including cleaning, laundry, fetching water, cooking and childcare – was allocated unevenly and fell by rule to wives. Marriage structured the equation between women and domestic work as women contracted the responsibility for domestic work once married. Whereas within polygynous households, wives organised housework in turns (of one, two or more consecutive days), in non-polygynous households single-wives could share their domestic duties with other available women, most likely daughters-in-law and sisters. Cyclical power structures consisted mainly in the fact that a wife would take over her mother-in-law’s domestic duties and be eventually relieved, years later, by her own sons’ wife/wives (and be substantially disadvantaged otherwise). Reflecting this cycle, interviews emphasised a tripartite differentiation of women.

First, there were older married women in their fifties (first, second and third wives), ‘freed from domestic work’, which had been taken over by their daughters-in-law. Their powerful status in the household had some mirroring in the workplace where they held lower-management positions. ⁶ Indeed, the availability of daughters-in-law constituted the most widespread solution to the domestic work burden. For a 33-year-old, unmarried woman juggling domestic and factory work the greatest problem was precisely the lack of a brother’s wife ‘my mother did not have a son that could marry and give us a sister-in-law to help us with the domestic work (. . .) if I marry and leave home she will be in great trouble’.

The second group included younger married women with young children that juggled factory and domestic work. The bulk managed by sharing domestic work with sisters and sisters-in-law. Those without this help were left unemployed because they were unable to take employment in the more distant farm. Two others were living with their brothers and their wives waiting to join their husbands in the marital household. Generally, for this group the contradiction between domestic and factory work was most intense: some had to give up factory work, some could not join the marital household, and some could juggle the two thanks the contingent availability of other women.

The third group comprised younger, unmarried women not yet invested by the full responsibility of domestic work. These girls undertook lighter domestic work supporting their mothers when not in school, but would be initiated to it very young as a core component of their upbringing and skill development (as future housewives). For example, some of them would work off-season as housemaids to support their parents with their wage.

The disciplining of women as housewives was not limited to unpaid domestic work. A crucial element of this consisted in patrilocality as the spatial disciplining of women within a household different from their original one. In consonance with patrilocal rules, once married, women had to transition from their parents to their husband’s household. This was invariably a traumatic transition where women abandoned the immediate support of their parents to negotiate a position in a new household, itself a theatre of established power relations. When entering a large household, new housewives could automatically become daughters-in-law and had to negotiate their tasks with their mother-in-law, and possibly other wives, sisters, aunties-in-law, etc. Several pointed at this transition as the most difficult in their lives. An 18-year-old woman still living with her mother dreaded that moment: ‘I will be very sad when I will have to leave my mother to join my husband home once married. . . but as a woman I have to marry, it’s a necessary evil’. For one, the relationships with the mother and sister-in-law had been so bad as to push her to divorce. Explicit forms of disciplining echoed from accounts of arranged marriages from older women, and from all interviewees praising patience and obedience as crucial female qualities representing standard female behaviour for ‘husband retention’.

A third major element structuring women’s position in the household in relation to men and other women was polygyny. Polygyny, or its mere possibility (in smaller households), represented perhaps the most piercing instrument of disciplining female identity and behaviour towards complex patterns of simultaneous cooperation and competition. Despite evidence in the literature of women asking for polygynous solutions (Gning and Antoine, 2015), one woman emphasised: ‘The majority of women do not like to find themselves with a polygynous husband. Marriage, I think, is not a matter of choice but of luck’. Women cooperated in domestic and paid work, while competing on performance to build their irreplaceability. This pressure ran strongly even for single-wives, never entirely free from the threat of becoming a co-wife: ‘I don’t want a co-wife. I can do all the domestic work by myself, without the help of anyone, even if it’s very hard’. As argued earlier, the prevalence of polygyny and its construction of women as structurally or potentially insufficient – hence to be juxtaposed and/or replaced by other women – worked as an incentive to intensify their work and prove their worth. This included the ability to avoid being an economic burden on their husbands prompting women to combine domestic work with paid, invariably low paid, and low skilled, informal and casual work.

The relationship between domestic and paid work highlighted two important elements. First, paid work was generally undertaken after fulfilling domestic and fields tasks. All women interviewed engaged in one or more forms of paid work. These included petty trading in fruit and vegetables clothes, or cosmetics; collecting and trading of gravel; preparing/selling breakfast; hairdressing for ceremonies; tailoring; paid domestic work (for young unmarried women); and finally wage employment in packaging centres, the most widespread full-time, yet seasonal, type of work available in the area. Yet, their involvement in paid work varied tremendously according to their relative position within the household, and their ability to delegate/share work with a co-wife, a sister, or a friend. So, both younger and older women ‘freed from housework’ could become a candidate for factory employment more easily than younger married women, whose conflict between domestic and factory duties was more intense and contingent on the presence of other women.

Second, domestic duties did not relieve financial pressures for women, who still needed to actively contribute to the depense quotidienne. Put simply, in adulthood domestic work was a prerequisite but rarely a substitute for paid work, hence women had to juggle the two. If at one level this pressure depended on the broader household composition and the different sources of revenue flowing in, on another level it ‘stuck’ to women as an underlying compulsion to demonstrate their irreplaceability.

How did women respond to these complex circumstances? All the women interviewed were engaged in the constant negotiation, re-working, and opposition to ascribed roles and duties. Women would mobilise their network to contact employers, skilfully combine tasks and undertake petty-trade ventures. Their paid jobs appeared frustrating (physically hard, time-consuming and poorly paid) and simultaneously appealing (a source of independence and a tool to break domestic routines) a means ‘to leave you worries behind’ as several repeated.

To juggle and individually counter their domestic burden, respondents combined, organised, allocated, and managed domestic duties in various ways, shaping a complex division of labour, itself riddled with the power relations sketched above. Opposition to domestic work emerged from several young women who refused a number of suitors and actively delayed their marriage as long as possible to avoid renouncing wage work. A divorced woman was not inclined to marry again: ‘I’m not ready to engage myself again because if I marry I could lose my job if my husband prohibits it (. . .). It’s difficult to find a supportive husband who lets you go to work rather than spend most day at home’.

All women in the sample retained relative control of their wage, though factory employment worked both for and against prevailing power relations in the household. For example, older women could finally break free from their domestic duties while reinforcing their influence vis-a-vis younger ones, hence increasing the burden of unpaid work on the latter. Young unmarried women could increase their independence from parents, though at the expense of school attendance generally reinforcing girls’ broader tendency to abandon school before boys. Finally, for wives the wage meant a degree of independence from their husbands, though most of the income earned would be spent on children’s sustenance, sometimes saved for the end of the wage season, and rarely invested in any other economic activities. So, wages gave them some autonomy without providing fuller economic independence from the household.

Perhaps the most manifest and collective antidote to different forms of household oppression, consisted in female participation in informal rotating credit groups, called tontines, and more formal institutions known as groups de intérêt économique (GIE). ⁷ Weekly, fortnightly, or monthly tontine meetings allowed both saving and socialising. Almost all women participated in one or more tontines: some were based on a village area, some on maternal kinship or religious affiliation, others were work-based. All worked by collecting regular, individual contributions from participants and providing them with a lump sum on a rotation basis. This sum was used as every-day extra-money, for social functions (marriages, baptisms, funerals) and sometimes to finance other small economic activities. For all interviewees tontines represented a financial help and a break from their routines, a space other than the household/workplace where they could socialise with other women.

Overall women’s ability to ‘get by’ with internal management, cooperation and competition ⁸ was itself the product of an overarching system magnifying their domestic work and care. As one put it: ‘Domestic work is multiple and painful, it’s a shackle in my life’. Although diverse and improving some aspects of their lives, female responses did not substantially challenge combined forms of disciplining, which constructed them as already precarious, contingent, and transitory (paid) workers. The requirement to solve domestic work before undertaking paid work had two simultaneous consequences: first the internal segmentation and unequal distribution of paid and unpaid work among women, and second the further fetishisation of domestic work for some of them. Indeed, behind each woman working at the factory, there would be one or more – a daughter, a daughter-in-law, a sister, a co-wife, a niece – taking over her domestic workload and therefore working unpaid.

Exploitation, control and resistance in the factory

Like households, factories reflected a fairly rigid sexual division of labour: women were employed largely for sorting, packaging, weighing, controlling and cleaning whereas men were employed in loading and unloading, transport, and higher management. The supply of labour from different villages was organised in (highly variable) different shifts, the working day including eight, ten or more hours. Work was generally unpredictable and depended on the influx of produce to the factories. A number of women constituted the reserve: they would go to the factory every day hoping to take someone’s place or respond to a surge in labour needs.

Work was accessed by registering their ID and receiving a badge and a uniform. Access to work was regulated more informally by sourveillantes, i.e. women simultaneously responsible for controlling packaging activities and supplying workers daily. One of the two firms had been systematically less reliable in paying workers and securing work. During fieldwork this firm had closed the packaging centre for a combination of reasons including a generally poor harvest across the sector. Interviews highlighted a three-fold system of exploitation centred around piecework, payment irregularities, and the externalisation of costs of re/production. These are sketched below.

Women sorting and packaging French beans were paid by the number of crates and bags packed, whilst a handful of surveillantes and peseuses (women weighing crates) were paid daily. Payments were either weekly or monthly, depending on the firm. Piecework exploited workers essentially in two ways. First, it drove the intensification of work: all interviewees would work as fast and as long as possible to earn more. Second, piecework allowed employers to offload a number of costs and risks to the workers themselves. Each crate was carefully controlled; crates not satisfying the inspections were not paid, resulting in a loss for the worker. Moreover, for all women work was more difficult, stressful and poorly paid at the beginning and end of the season, i.e. when the quality of the produce was poorer, the number of beans to be discarded larger, and handling lengthier and more difficult. When asked what she liked least at work, a woman claimed: ‘The lack of beans. If there are enough beans I am ready for everything’.

Interviews also revealed instances of delayed, irregular, and failed payments. While women recalled these problems as ubiquitous in the past, systematic delayed payments persisted in one firm. Here delayed payments often went hand in hand with cheating: ‘every time we do a crate we get a token. . . at the end of the month when we receive the pay we compare it with the job done and notice a gap and even if you protest you obtain nothing’. Alongside these more or less fraudulent practices, the most pervasive form of exploitation consisted in the systematic externalisation of the cost of re/production to the workers (Dunaway, 2014; Mies, 2014; O’Laughlin, 2013).

The largest cost of labour reproduction consisted in seasonality, i.e. the absence of factory employment for most of the year. By working at the factory only four/five months a year, women could contribute money income to their personal and household reproduction only partially, the rest of the time relying on other side activities and/or experiencing significant hardship. Moreover, other major reproduction costs – e.g. sickness, maternity/care leave – could incidentally fall during the working season. Interviews revealed the deeply intermittent nature of female employment: women typically quit work when entering their husband’s household, had children, or cared for a sick parent, a sibling, or a sister with a new-born. The possibility of quitting and re-entering factory work, after one or more seasons, emphasised broad employment availability as well as the strong interchangeability of workers: replaceability at work was therefore a double-edged sword for women.

On the other hand, behind the paid work of each female employee lay the unpaid work of a mother, a daughter, a daughter-in-law, a sister, or a niece: the daily reproduction cost of female workers at the bottom of the chain fell on this further invisible army of women. However, some of the off-loading of re/production costs on the part of the employers at times made this work more visible. For example, every day, women at home would bring lunch directly to the factory or the company bus for collection. Workers without this ‘back up’ could not take employment far from home and would not be able to work a full day because they would have to fill the cooking gap. Here, the (gendered) home made up for the cost saved by the employers.

Overall, these different forms of exploitation relied on a multi-layered infrastructure of worker disciplining. Indeed, to cater for seasonality, piecework, informality, and the gendered worker, workers were subject to work, time, body, and gender disciplines. Work discipline was very much internalised by each worker through piecework and the associated intensification of work. Interviews revealed the workers’ complex juggling and balancing of speed and quality pressures, invariably attributed to one’s skills. Disciplining also relied on a system of more manifest rules and threats based on the control over work. Controllers walked up and down the tables checking sorting, packaging and appropriate worker behaviour. Misbehaviour or poor work were met with warnings and eventual dismissal. Time discipline emerged both from female internalisation of speed due to piecework as well as maximum flexibility in terms of the working day. Women described an extremely variable range, including very long, intense days, very long and not very intense days with a great deal of (unpaid) time waiting for the produce to arrive. From their perspective, what was invariable was precisely the unpredictability of the working day ruling their time and lives. Against this maximum flexibility on their part, there was very little flexibility on the part of the employers, e.g. women late for work after the lunch break would be excluded from the afternoon shift and lose the whole day’s pay.

Body discipline included lengthy hours of repetitive tasks either standing or seated, or working in cold rooms. Women reported problems to their necks, backs, knees and eyes. All stressed a strict body discipline, requiring covering their hair, strict hand hygiene, the ban on any jewellery or beauty products, the prohibition on eating, chatting, or using the phone during work. Whereas part of these rules stemmed from hygiene requirements, the demands for silence, patience, and immobility were considered part of their female qualities that included great tolerance, docility, and nimbleness. All women had internalised sorting and packaging as a ‘natural female job’ as opposed to ‘male jobs’ which required physical strength and vigour. Indeed, the division of labour inside the factory had been naturalised: ‘This work is not a male thing. They underestimate it thinking it does not pay much. It’s a job that requires patience’. Crucially, patience and subordination were praised as an instrument for job retention, vis-a-vis innate male propensity to be more hot-headed, audacious, daring, and consequently more susceptible to losing their jobs. Thus, the society-ascribed quintessential female qualities matched the requirements within the workplace enhancing once again the continuity between the household and the factory.

When encountering problems women responded predominantly with individual strategies. Their first action was ubiquitously to bring their complaints to their supervisors, mostly to the surveillantes. Occasionally, the problem was voiced as a small group. Some were good at defying some rules, especially in terms of work access: several used different strategies to conceal their young age, including complicity with a supervisor, hiding within the morning crowd entering the factory and sitting on boxes to appear taller. In most informal settings women shared the work to combine their household and factory jobs. As a last resort, when faced with repetitive delayed or failed payments, women responded by quitting the job.

Knowledge of, and participation in, more organised collective action was patchy. A few women narrated past struggles and work stoppages for salary increases. However, for most respondents strikes were generally perceived as within a male sphere of action, as something that men would typically do, and pay for. Instead, resilience was often invoked as a safer, smarter and more long-lasting strategy, which did not risk one’s job. Almost all acknowledged the meagre salaries received but would invariably conclude with a ‘on fait avec’. Overall, overt, collective resistance was male, while perseverance was female. Here the construction of gendered workers involved their elaboration of resilience and patience as a form of resistance itself. In short, by leaking into the sphere of resistance, the construction of gender worked again to make women capital’s ‘perfect workers’ (Mies, 2014).

Conclusions: Structural relations between households and factories and challenges for class struggle

This article has focused on how exploitation and the long-standing feminisation of labour at the bottom of supply chains operate in the context of Senegalese export horticulture. Here, the ‘construction of gender as a set of context specific meanings and practices’ (Bair, 2010: 205) is manifest in ascribing women as the peasant reproducers, a role resulting from the articulation of pre-colonial and colonial legacies, religion, the state and capitalist developments. To explore these processes – indeed to understand how cheap labour is constantly reproduced – the article advocated the broadening of labour control analysis through the systematic engagement with the sphere of social reproduction, in particular the analysis of household relations as structurally linked with workplaces. This was done by investigating the multi-faceted disciplining of women at home and in factories and its relation to exploitation.

In rural Senegal, the women interviewed are disciplined as quintessential reproducers and carers; this role determines – but never substitutes – their secondary role as income-earners (Elson, 1999). Their role as wives precedes their role as wage earners, thus women become permanently transient and precarious members of the working class filling up the lowest jobs. Within households, women engage in relations of cooperation and competition as polygyny, actual or potential, drives competitive pressures and the threat of replaceability. This translates in turn into their asymmetrical, relational, and cyclical juggling of unpaid and paid work. Thus, female disciplining serves the reproduction of both capital and labour: as wives and mothers they reproduce labour while their transient and nimble paid work reproduces capital.

In factories exploitation centres around piecework and capital’s ability to profit from delayed, at times deceitful, payments and the systematic offloading of re/production costs to workers. The latter is met by the household in multiple ways, by women workers who abandon paid work to care for someone and by other women who fill the domestic duty gap. In other words, the cyclical alternation of women at home pays for externalised social re/production costs ‘that reduce the wage bill of the capitalist’ (Dunaway, 2014: 60).

Overall, in this context the constant supply of cheap labour stems from the interaction between the household and the workplace. Capital reproduces cheap workers through a complex interplay between exploitation and disciplining that subordinates workers’ bodies and identities to maximum value production at minimum costs. The household reproduces capital through its internal political economy and invisible army of female domestic workers that make exploitation possible. Women move across different spaces and regimes of control, the household and the firm, but rarely break free from either, since the two spaces of control are dialectically related. The household produces and reproduces a reserve army of labour, especially in the form of cheap female labour, some of which capital employs at very low cost by paying well below the cost of their reproduction. This increases female cheapness and reproduces the existence of the large household as the stereotypical African safety net (Therborn, 2006), simultaneously against, and in support of, exploitation. The latter is far more complex than partial access to employment and cannot fully be resolved via more employment. In other words, it is not (exclusively) the lack of ‘proletarianisation’ marginalising women, but rather how capitalism incorporates and interacts with the household and contributes to shape the division of labour therein (Federici, 2012; Mies, 2014).

The relation between production and social reproduction also stretches to the sphere of resistance. Indeed, if disciplining women to a reproductive femininity represents a form of class struggle from above, what is left of class struggle from below – i.e. workers’ class struggle? Women engage in different strategies at home and in factories as they ‘struggle as women’ and ‘struggle as workers’ (Elson and Pearson, 2011). Some of these are congenial to capital: resilience to exploitation manifests above all female internalisation of disciplining. Opposition emerges when women reverse their ascribed roles of obedient and submissive wives, daughters and mothers. So, women disobey when breaking rules, e.g. by lying about their age and delaying/refusing to marry. Most instances are however individual and crucially extend their wage labour life, and the exploitation that accompanies this: they transfer them to another terrain of control rather than undermining it. Similarly, by participating in self-help saving groups, women organise alternative networks and relationships to the households while socialising, evading work routines and making the most of the money earned. Yet, while this manifests strong individual and collective agency, these groups work through the financial self-disciplining of women rather than challenging low wages and capital accumulation. This suggests that separate forms of struggle reproduce and reinforce the (artificial) distinction between women as reproducers and wage workers hence obscuring rather than exposing the structural relations between the two (Barca and Leonardi, 2018).

Ultimately, the case of Senegal unveils some key aspects of how exploitation operates at the bottom of the supply chain in the global South. Here capital and the state reinforce relations between households and factories/farms because ‘neither the market economy nor the state are willing to commit to reproducing working class households’ (Naidu and Ossome, 2016). The ‘legendary’ African extended household is the reproduction regime that supplies cheap labour for export food production by assuming the costs of that cheapness. This cost is borne ultimately by invisible, shifting, armies of women scrambling to manage the inherent contradiction between production and social reproduction (Fraser, 2016). This contradiction is more intense because rooted in North-South relations: the supply chain is the racial mechanism that transfers the social cost of low wages to the South where the extended, peasant, polygynous household, and its gendered classes of labour are the result of global relations of production, rather than a lack of it.