From the shop floor to the kitchen table: the shifting centre of precarious workers’ politics in South Africa

The labour movement in the global South is in the midst of a prolonged crisis. Over the past few decades the financialisation of capital, the globalisation of production and the rise of ideologies of flexibility in the workplace have undermined the formal wage workers who are the traditional base of organised labour. This crisis has sparked a wide-ranging debate on the future of the labour movement. Many scholars and activists have turned to the question of labour movement revitalisation, searching for strategies and tactics that might allow existing unions to incorporate the growing ranks of unorganised and precarious workers (Bonner and Spooner 2011; Milkman and Voss 2004; Turner 2005; Von Holdt and Webster 2008). While these authors recognise the great challenge posed to unions by the rise of precarious forms of work, they remained convinced that, as Bonner and Spooner have put it,

In contrast, other observers have taken the position that labour unions are unlikely to regain their role as the primary organisational home of the working class. In this view, the erosion of formal wage work has undermined the basis of the broadly shared material interests that defined traditional labour politics. As a result, unions are seen as a form of political organisation whose time is past. Guy Standing, for example, has argued that because unions were built to represent workers in a specific type of formal work arrangement, ‘[p]rogressives must stop expecting unions to become something contrary to their functions’ (2011, 168). In Manuel Castells' words, the unions of today are doomed to ‘[run] behind the new society, like dusty flags of forgotten wars’ (2004, 420).

In recent years, a third position has arisen in this debate which can be summarised as a focus on what Ching Kwan Lee and Yelizavetta Kofman call a ‘politics of precarity’ (Lee and Kofman 2012). This position starts from the observation that collective action around issues related to work has not disappeared, and is unlikely to, despite the rise of precarious forms of employment. However, in contrast to the labour revitalisation literature, this work does not focus on precarious workers from the perspective of existing unions, but instead treats them as important contemporary political actors in their own right. These authors take as a given that precarious workers will continue to struggle, but not necessarily through the same organisational forms and over the same issues as do workers of the traditional labour movement. The declining relevance of the ‘traditional politics of labour’ (Paret, forthcoming) to the majority of the world's workers requires a shift in focus to the emergent politics of precarity for scholars and activists who want to understand the future of economic struggles.

Although this literature is relatively new, it has begun to produce important insights into the new forms that precarious workers’ politics are taking around the world. Jennifer Chun (2009) has highlighted the way in which marginalised workers in South Korea and the United States have turned to ‘symbolic leverage’ as traditional forms of workers’ power have been eroded. She argues that marginalised workers often attempt to ‘[redirect] the site of struggle from narrowly defined workplace disputes to public contestations over values and meanings' (173), making appeals which are based in ‘moral and cultural understandings [as much as] economic calculations … ’ (7). Marcel Paret finds a similar style of claim making among precarious workers in Gauteng, South Africa whose demands reflect a ‘politics of recognition’ centred on a struggle for dignity and social worth rather than simply workplace-based demands (Paret, forthcoming). Rina Agarwala, studying home workers in India, has shown that, in contrast to traditional labour organisations that put demands to employers, precarious workers tend to see the state as the actor responsible for providing for their well-being (Agarwala 2013). Similarly, Matteo Rizzo's study of taxi drivers in Dar es Salaam, Tanzania, shows the role that appeals to the state can play, not only for questions of general social protection, but as a lever of structural power which precarious workers can use against their immediate employers. Rizzo details the example of Dar es Salaam's drivers who lobbied the government to amend laws so that negligent taxi owners, rather than the drivers they employ, would be held responsible for traffic violations associated with unsafe taxi vehicles (Rizzo 2013). Lee and Kofman, summarising the insights of this emerging literature, conclude that,

From politics of labour to a politics of precarity

This analysis of collective action by precarious workers in the global South is essential to understanding the changes taking place in class politics with the erosion of formal wage labour. However, the emerging literature on precarious politics suffers from the limitation that the examples of collective action on which it focuses remain, in almost all parts of the world, relatively isolated and uncommon events. The vast majority of precarious workers are not organised and do not make collective demands around issues of work. It is still unclear what the implications of these innovative, but still numerically marginal, precarious workers movements are for the unorganised majority of precarious workers in the global South. Do these new forms of organising and making demands point towards the possibility of a revitalisation of a labour movement that could represent the interests of the excluded majority of contemporary global capitalism? Or will such experiments remain rooted in a small section of the swelling ranks of precarious workers?

A more complete understanding of how the rise of precarious work has transformed workers’ politics cannot come only from an analysis of scattered examples of precarious workers' organisations. Instead it must proceed from an analysis of the social and economic conditions of precarious workers as a whole, which may (or may not) provide common interests and experiences around which a precarious politics could form. In this way, the concept of a politics of precarity should mirror the concept of a traditional politics of labour which it is critiquing.

The concept of a ‘politics of labour’ conveys the idea that workers in capitalist economies share certain material and political interests by virtue of their position in the production process. This idea is not drawn only from an observation of workers’ organisations and collective action. Instead it is rooted in a set of clear assumptions around workers and their experience of work. The ideo-typical workers who advanced the traditional politics of labour were assumed to be fully proletariansed, that is, their livelihood depended on their ability to earn a wage. Solidarity and commonality of interests were assumed to result from the shared daily experience of a common workplace. In most cases of sustained traditional labour movements, some form of social contract was assumed to govern the relationship between employers and workers, setting limits on the terrain of labour–capital conflict.

It is reasonable to assume that as the validity of these assumptions has been eroded, the political interests of the working class have also changed. However, understanding the emerging politics of precarity is not as simple as turning our attention from the common experiences of traditional workers to those of precarious workers. The term ‘precarious work' is used to describe a broad and diverse range of experiences, from wage workers in outsourced, part-time or temporary arrangements to the unemployed and self-employed poor who make up what Michael Denning (2010) calls the ‘wageless’ segments of contemporary globalised capitalism. In order to understand what, if any, political interests this broad range of workers shares it is necessary to ask what the common experiences are of precarious work. This article takes up this question in a specific place, contemporary South Africa. The article draws on nationally representative household survey data combined with interviews and observations from the author's field work among precarious workers in both rural and urban sites in South Africa. It asks the question, are there common experiences among the diversity of precarious workers that point towards some general politics of precarity in South Africa?

The findings of the article both confirm and add specificity to the emerging concept of a politics of precarity. The data from South Africa show that precarious workers have complex economic lives, relying on a combination of diverse income sources including, but not limited to, their own wage in order to gain a livelihood. The primary site through which incomes are combined and livelihoods are produced is the precarious workers' households. As a result, these workers' material interests are centred on their household livelihood strategies rather than their workplace. For many of these individuals, their primary identity is not that of precarious worker but rather that of a family or household member. Increasingly, class interests and identities are constructed not on the shop floor but around the kitchen table. This shift has important implications for attempts to build broad-based organisations or coalitions among precarious workers. What unites these workers is not their experience of work, but their experience of precarity which requires diversified household-centred livelihood strategies.

Precarious households: the hidden abode of reproduction

Traditional conceptions of the politics of labour have been rooted firmly in the workplace. Workers' material interests are thought to be shaped by their experiences in that hidden abode of production. As ideo-typical formal wage work has been eroded, workers’ ties to the workplace have been loosened and they have increasingly begun to rely on social and kinship networks in order to gain a livelihood. The household, what Bill Martin and Mark Beittel have called ‘the hidden abode of reproduction’ (1987), has become increasingly central in shaping precarious workers’ material well-being. The importance of the household is neither new nor exclusive to precarious workers. As Joan Smith and Immanuel Wallerstein have argued:

However, the rise of precarious work has increased the centrality of the household, shifting a larger portion of the burden of social reproduction from the labour market to the household.

Given its importance to precarious workers’ livelihoods, the household is a key site in which to think about the politics of precarity. For precarious workers the experience of the workplace varies widely, ranging from something resembling a ‘standard employment relationship’ to informalised sites of self-employment to periods of open unemployment in which there is no workplace at all. The vast majority of precarious workers do, despite this variation in workplace experiences, situate their livelihood strategies within a social network that can be described as a household. Households are typically centred on a common place of residence, but household connections, in the sense of Smith and Wallerstein's definition, can extend beyond a specific location, often, for example, stretching across spaces of migration.

In the analysis presented below, precarious workers are situated within households. This is not meant to romanticise the household as a site of mutuality and cooperation. As has been well documented, households are sites of conflict and subject to unequal power dynamics. The increasing reliance on households as sources of well-being may be exacerbating such conflict (Mosoetsa 2011). However, despite this conflict, households are the central site through which workers’ livelihood strategies are carried out, as the data below will demonstrate.

Conceptualising households in the broad sense that Smith and Wallerstein have presented is easy enough. A much more difficult task is accurately analysing the dynamics of households on a large scale. There is a wide range of household survey data available, but much of it suffers from limitations which are especially important to consider when analysing precarious workers’ households. One key issue is determining how to define a household member. A household survey that does not, for example, measure connections between urban migrants and their rural homes will severely distort the actual picture of a worker's economic situation. This is particularly true for precarious workers, whose households and livelihood strategies often extend beyond a single residential location.

A second problem is measuring all sources of households’ income through a survey. Households are complex sites of cooperation and conflict in which multiple types of income are combined and/or produced. Smith and Wallerstein defined five types of income which households utilise in order to gain a livelihood over time: wages, market sales, rent, transfers and subsistence (1992, 7). The first four of these can be measured with a fair degree of accuracy and reliability through detailed household surveys. Subsistence income, by contrast, is very difficult to quantify, yet it is essential to the livelihoods of almost all households. Subsistence income is especially important for precarious workers since their other forms of income tend to be low.

This article combines two sources of data in order to analyse the households of precarious workers in South Africa. First, it uses quantitative data from the National Income Dynamics Study (NIDS), a data-gathering project by the Southern African Labour and Development Research Unit (SALDRU) at the University of Cape Town. The NIDS is a large-scale, nationally representative survey of 8040 households containing just over 32,000 residents. Although the NIDS is a panel survey, conducted bi-annually since 2008, the data used in this article come only from the most recent wave, conducted in 2012 (SALDRU 2013). The NIDS was specifically designed to address some of the shortcomings in existing household data. It uses a very broad definition of household, counting anyone who lives in a common physical place at least 15 days out of the year as a household member. The survey also includes a number of questions that collect information about additional economic ties that may exist with individuals who are not counted as household members. However, despite these advantages over previously existing household surveys, the NIDS data do not allow an accurate assessment of subsistence income. In fact, given its complexity, a comprehensive and generalisable measure of subsistence income would be virtually impossible. In order to illustrate the way in which subsistence income shapes the material interests of precarious workers, this article uses data from the author's field work among precarious workers in three rural sites in Mpumalanga and KwaZulu-Natal and from across multiple urban sites in the province of Gauteng. This field work, conducted in 2010 and 2011, involved interviews and ethnographic observation focused on households’ economic situations with an emphasis on social connections and subsistence incomes.

Conceptualising and operationalising precarious work

Despite, or perhaps because of, its ubiquity in contemporary debates, there is no broad consensus on the definition of precarious work. Standing (2011) identifies precarious workers as a distinct and growing class within the labour force who lack forms of security associated with the traditional working class. By contrast, Franco Barchiesi, in his study of precariousness in South Africa, notes that ‘the vulnerability and precariousness of employment … are not confined to the lack of formal jobs … . Even many workers enjoying the protections of unionisation earn wages that often barely cover basic necessities’ (2011, 75). Barchiesi's findings seem to confirm Ronaldo Munck's scepticism of the usefulness of the concept of precarious work, especially in the global South. Munck laments that ‘[t]here is little cognisance [in the literature that] the type of work described by the term “precarity” has always been the norm in the global South’ (2013, 752).

Indeed, Munck's critique could be extended beyond the global South. In a sense, the vast majority of workers in a capitalist system are in a precarious situation, in that they depend on the labour market in order to gain all or part of their livelihood. It is this inherent precarity that is the underlying reason for labour laws which provide workers protections from the market by, for example, banning arbitrary dismissal, guaranteeing minimum wage or working conditions etc. While we may debate the novelty of this situation, it is undeniable that in contemporary global capitalism a significant portion of workers work outside of these protections. This precaritisation of work has been driven by two converging processes. On the one hand, employers have found ways to exclude previously protected workers from the framework of labour protection offered by the law. This is what Jan Theron has called ‘informalization from above’ (Theron 2010). At the same time there has been an ongoing ‘informalization from below’ (Ibid.) driven by the masses of individuals who are excess to the labour requirements of global capitalism and who are forced to make their own work in order to survive. Even if we reject Standing's argument that these workers constitute a distinct class, we can ask how the precarious reality of work shapes the politics of the contemporary labour force.

In the quantitative analysis presented in this article, precarious workers are divided into two broad categories. The first category contains primarily workers who are the product of informalisation (or precaritisation) from above. These are workers who have a regular wage job, but one which lacks the full range of legal protections available to workers under South African labour law. These workers can be further divided into two sub-categories. First are those workers in jobs that should be covered by the law, but that, in actual practice, do not meet the minimum legal requirements. This group of workers will be called unregulated wage workers. In South Africa, these workers are both the largest group of precarious workers and the most difficult to quantify, since their extra-legal work arrangements cannot be easily measured through survey techniques. The NIDS survey does provide some questions which make it possible to determine if the respondents’ conditions of employment adhere to basic South African labour law. For example, the Basic Conditions of Employment Act (BCEA) requires that all workers who work more than 24 hours per week must be provided a written contract. Therefore, those workers who report that they work more than 24 hours a week but do not have a contract can be understood to be working in an unregulated job. Similarly, full-time employees must have contributions to the national Unemployment Insurance Fund (UIF) deducted from their pay check. Those that report that they do not have UIF deducted from their checks are in jobs that contravene the BCEA.¹ While it does seem likely that such small violations of labour law are a reasonable proxy for the larger forms of insecurity that define precarious employment, these types of questions are unlikely to capture the full range of unregulated wage work that exists in South Africa. Therefore it should be expected that unregulated wage workers are undercounted in the data presented below.

These unregulated workers are in jobs that operate outside the legal framework of South African labour relations. There is another group of workers who have regular wage jobs which do adhere to the law but still qualify as precarious because their employers use legal means to circumvent some of the protections provided by the legal framework. This category includes outsourced workers, workers who are employed on a fixed-term contract and workers who are employed part time. For all of these part-time/contract wage workers, various aspects of the law can be circumvented (Theron 2003). Most significantly, they are easier to dismiss. In contrast to unregulated workers, the NIDS data should provide a relatively accurate measurement of part-time/contract workers.

Both unregulated and part-time/contract workers have something that looks like a regular wage job, even if it is insecure and/or of limited duration. The second broad category of precarious workers, which mirrors Theron's concept of informalisation from below, consists of those workers who have no regular employment and instead rely on low-income survivalist activities. As with regular wage jobs, workers in survivalist activities can be divided into two sub-categories. The first category includes individuals who rely on self-employment. Of course, not all of the self-employed are in precarious work. For example, successful business owners or wealthy consultants should not be grouped together with street hawkers or individuals offering haircuts in their backyards. For this article, self-employed individuals are only counted as precarious workers if they earn R3100 per month (∼$310) or less from their business.²

The second type of precarious workers who rely on survivalist activities are those individuals who work as casual workers, day labourers or pieceworkers. Of course, these casual workers do earn wages. However, unlike the regular wage workers described above, they do not have consistent and reliable access to wage income. This includes individuals who find day work on the roadside or by approaching work sites. It might also include individuals who are occasionally hired by relatives or neighbours for specific tasks, but on an inconsistent basis.

A final small group of workers who are counted as precarious are those who report in the NIDS survey that their main economic activity is helping a family member with his or her business. This type of work begins to blur into other types of unpaid household labour (such as childcare and food preparation) which are not included here as forms of precarious work. However, these individuals are included because their unpaid work is oriented directly towards market activities.

There are three groups of workers who should be mentioned who are not counted as precarious workers in the data below, one of which is omitted for conceptual reasons, the other two owing to the limitations of the NIDS data. The first and largest group is the unemployed. The unemployed share many material and political interests with precarious workers and should be considered important to the overall politics of precarity. However, the central argument of this article, that precarious workers rely primarily on household-centred rather than workplace-centred livelihood strategies, is obvious in the case of the unemployed. If the unemployed are considered alongside the precariously employed workers who are the focus here, the arguments of the article can only be strengthened.

The second group is excluded from the analysis below because of limitations of the NIDS data. When the NIDS fieldwork was undertaken some household members were not available to be interviewed, often because they were migrant workers who lived and worked much of the year away from their ‘sending’ household. In these cases other household members were interviewed as proxies for the unavailable respondents. These proxy interviews asked a simplified version of the full adult survey. The proxy questionnaire does not make it possible to distinguish workers who are unregulated or part-time/contract workers from ‘non-precarious’ wage workers. Since these proxy respondents are primarily migrant workers they are probably more likely to be in precarious work than a randomly selected worker. Yet, without further information this group of workers, who constitute about 11% of all employed workers, cannot be reliably classified. They are described in the charts and analysis below as unknown employed workers.

The final omitted precarious workers are the heterogeneous group of individuals who work in what would be conceptually defined as precarious jobs, but which the data provide no way of identifying. For example, individuals who had been subcontracted out from their primary employer would not be captured as precarious if their contract was full time and not of a limited duration. Other workers might be subjected to conditions of employment that do not adhere to South African labour law, but if they had a written contract and had UIF deducted from their check, they would not be identified as precarious. Because of these limitations the group of precarious workers analysed below should be considered a conservative estimate of the overall population of South Africa's precarious workers. Although the terms ‘precarious workers’ and ‘non-precarious workers’ are used below for the sake of convenience, it would be more accurate to think of the groups discussed as those whose job can or cannot be identified as precarious by the NIDS data.

Using the above methodology, 42% of South Africa's employed labour force can be identified as precarious.³ Of households that have an employed member, 48% contain a worker who can be identified as precarious in the NIDS data. Considering the conservative method of counting precarious workers used here, it is safe to assume that precarious workers' households are the modal households among the employed in SA. Figure 1 presents a breakdown of the employed labour force as a whole (left pie) and of precarious workers by type of work (right pie). Just under two-thirds of precarious workers are in regular wage work while the remaining workers engage in various forms of survivalist market activities.

Figure 1.

South African precarious workers by type.Source: author's calculations from the 2012 NIDS data (SALDRU 2013).

Precarious workers’ household livelihoods

Figure 1 shows the position of individual precarious workers in the labour market. However, because it shows nothing about precarious workers’ households, it obscures the full material reality of these workers’ economic lives. Table 1 shows the full range of income sources on which precarious workers’ household rely. Precarious workers’ households are any household that contains at least one precarious worker. In the table, households are separated into three strata depending on their position in the per capita income distribution. Income from work sources – including both wage work and self-employment – is divided into two categories, depending on whether the work that provides them would be classified as precarious or not. ‘Unknown work’ refers to the income earned by proxy respondents in the NIDS survey whose work cannot be categorised.

The table shows that wages from precarious work provide less than half of total income earned by all precarious workers in South Africa. For poorer precarious workers' households, the percentage of income contributed by wages from precarious jobs is less than 40%. This is evidence of both the low levels of precarious incomes and the diversity of income sources on which households rely. Even within precarious workers’ households, wages from formal jobs can be important to the households’ overall livelihood. This is reflected in the amount contributed by wages from ‘non-precarious’ work, but also by the amount contributed by remittances. Remittances and non-precarious work combined contribute over a quarter of total income to precarious workers’ households. While the structure of the survey does not make it possible to determine the original source of remittance income, it is safe to assume that a significant portion of remittance income is from non-precarious sources since precarious incomes are often too low to allow for remittance sending.

Government grants are another vital source of income, especially for poorer precarious workers’ households. Two programmes contribute the vast majority of these grants. The first is the state grant for older persons, known as the pension grant. This is a means-tested grant given to all South Africans over 60 who meet the means test.⁴ The second is the child support grant, a means-tested grant given to the guardians of children under 18 years old. These grants form the backbone of the country's social welfare system. In 2013 the South African Social Security Agency provided 15.6 million grants, including the two mentioned and a range of smaller grants (SASSA 2013). This amounts to just under one grant for every three residents of the country.

The next largest contribution to incomes is made by implied rent. Implied rent is an imputed value that reflects the amount that a household saves by not having to pay rent on its primary dwelling. Two categories of households ‘earn’ implied rent. First are those households in which a resident member owns the home in which the household resides. The second group is households who report that they neither own nor rent the home. These individuals could be living in a home owned by a non-resident family member or friend, or they could be ‘squatting’ or living in an informal dwelling on which no rent is paid. Of course, implied rent is, strictly speaking, a form of savings rather than a form of income. However, it is important to include here because this savings is an essential source of livelihood, in some cases the most important source for poorer households. The other income category combines a number of relatively small income sources. For wealthier households it is mostly income from investments. For poorer households it is mostly subsistence agriculture.

The presentation in Table 1 is useful for identifying overall trends. It clearly establishes the fact that wages from precarious jobs are only one of many income sources on which precarious workers’ households draw. However, because the table combines all income earned by all households in each stratum, it masks the fact that there is great diversity between households, even within a given income stratum. Individual households do not necessarily draw from all of these income sources at any given time. Figure 2 presents a better picture of the diverse types of households in which precarious workers live. In this figure precarious workers are divided according to the income source that their household relies on for the majority of its income. Thirty-eight per cent of precarious workers live in a household in which their own wage is the primary household income source. The remaining majority of precarious workers live in households in which an income source other than their own wage provides the most of the households’ income.

Figure 2.

What income source provides the majority of household income to precarious workers' households?Source: author's calculations from the 2012 NIDS data (SALDRU 2013).

For 25% of precarious workers, non-wage sources (including government grants, remittances and implied rent) provide most of the household's income. About the same percentage of precarious workers (22%) live in households in which the wages of other household members provide most income. The remaining 15% of precarious workers live in households in which no single income source accounts for more than half of total household income.

The data presented above demonstrate the degree to which precarious workers rely on income sources beyond their own work in order to gain a livelihood. They show that the material interests of precarious workers extend far beyond their own workplace. The data show the central role that the household plays as a site in which income is pooled from multiple sources. However, the income sources captured in the data presented here do not account for the full complexity of South African households’ livelihood strategies. Most precarious workers’ households are earning very low incomes. Even those households on the 75th percentile of the income distribution earn less than R2000 (∼$200) per capita each month. The median per capita household income among precarious workers is about R900 (∼$90) per month. To get by on such low levels of income, these households depend on support from social and kinship networks which cannot be as easily or reliably measured, but which is nonetheless vital to the livelihoods of precarious workers. This support is a form of what Smith and Wallerstein (1992) called subsistence income. The next section turns to the role of this form of income in South African precarious workers’ households.

Reading between the numbers: subsistence income and precarious workers’ livelihoods

Subsistence income is not only difficult to measure, it is also, as Smith and Wallerstein recognised, the ‘most confusing’ form of income to conceptualise. It captures the reality that ‘virtually every household produces some of what it requires to reproduce itself … ’ (Smith and Wallerstein 1992, 9, emphasis in original). Thinking about this self-produced output as a form of income conveys the idea that it is another resource that individuals and households utilise alongside wages, grants, rents and other monetary resources in order to gain a livelihood. However, subsistence income is more accurately measured not in monetary terms, but in terms of work. It is the outcome of the unpaid work that individuals depend on for survival and reproduction.

Some forms of subsistence work, such as growing food for household consumption, can be measured in monetary terms in the NIDS data. (These are mostly captured in the other income category in the tables and figures above.) However, the most common forms of subsistence work cannot be measured in the NIDS data. For example, the quotidian domestic tasks – such as cleaning, preparing food and caring for children, the ill or the elderly – are performed without compensation and are essential not only to South African precarious workers, but to the livelihoods of virtually all individuals in capitalist economies. For precarious households this unpaid domestic labour may be even more important than for wealthier households, given precarious workers’ low levels of other income.

However, subsistence labour is not only oriented towards immediate reproductive tasks. Unpaid household labour also plays a vital role in securing long-term livelihoods of individuals and families. Securing some level of stability in post-work years is an enormous challenge for precarious workers, many of whom struggle to get by even during their working years. This long-term oriented subsistence labour is important for understanding the economic lives of precarious workers.

One key to most individuals’ long-term livelihood plans is to have children who will work and be able to offer some support to their parents. Raising children is therefore both an everyday reproductive task but also a part of a family's long-term economic plans. The same applies to maintenance of a house or land which a family owns (or to which it has a secure long-term claim, in the case of areas governed by customary law). Having a house is one of the most important components of a household's long-term livelihood strategy. A place to live rent-free, especially in post-work years, provides a level of security that is difficult for many precarious workers’ households to achieve. For those families whose houses are situated in rural areas, often far from job opportunities, maintaining a house often means having at least one adult family member who forgoes the opportunity to travel in search of wage work. In this sense, the mere presence of an individual at a rural home over their working-age years can be thought of as a form of subsistence ‘labour’.

For both the everyday and the long-term varieties of subsistence labour, precarious workers are usually key providers for their households. This responsibility shapes these individuals’ roles and identities within families and households. Their precarious work is often a part of that strategy but is rarely the primary activity, especially for long-term livelihood strategies. This point is best illustrated by looking at an example of a South African precarious worker's household, members of which were interviewed in my field work in Mpumalanga province.

The primary respondent in this household is a married woman in her mid forties. She lives on the same property as her elderly mother-in-law, her children, her sister-in-law and her nephews. Her husband and his younger brother are also members of this household. Although they work most of the year in Johannesburg, they return to the home every December. The husband and brother-in-law work as wage labourers in the construction industry (in jobs that would be classified above as precarious) and also do independent construction work on their own. At the rural home the wife and sister-in-law take care of the children and the ailing mother-in-law, who receives a government pension grant. They also run a small shop out of their home which sells frozen fish to neighbours. In addition to attending school, the children take care of the animals and older ones help their father add buildings to their property when he is home for a few weeks each year. The property consists of two finished buildings and a number of buildings under construction, some of which are being used as dwellings while they are under construction.

Although this family has a low income, they have a more stable long-term livelihood strategy than the majority of their peers in their rural community. Being married, the couple will have access to two old-age pensions in their retirement. They also have a small but stable business, selling fish, which will supplement this income. They have animals which act as a safety-net store of wealth. Most importantly, they have a home that they live in rent-free, which saves them one of the major expenses most families face in retirement. If some of their children are able to find jobs that allow them to send occasional remittances, this couple could expect to be one of the more secure retired couples in their community after they finished work.

In a survey like the NIDS, the wife in this family would be counted as a self-employed precarious worker. However, her role in securing the family's long-term livelihood extends beyond the work she does selling fish from the home. Her most important subsistence contribution is the fact that by living at and maintaining the rural home she allows the family to continue their claim on the homestead. This role as the caretaker of the home is much more important to the long-term security of the family than is her income from her work, or any other precarious work income that she might be able to earn. She would not, for example, leave her home to take a precarious job in another location if there was not someone else to maintain the rural house. Income from precarious work is both insufficient and unreliable. It cannot be counted on to provide enough to secure housing for her family in retirement.

This might seem like an obvious insight. However, these details of precarious workers' actual lives and experiences are often ignored in debates about ‘organising’ such workers. Both the wife and the husband in this household are precarious workers, one self-employed (the wife) and one a casual or unregulated worker (the husband). However it is only when we look beyond these workers’ experiences of precarious work, situating their paid labour in the larger context of their household livelihood strategy, that we see the broad range of issues which affect their material interests. For these workers, access to land (primarily for housing) and access to pensions are as, if not more, important to their long-term well-being as are the precarious incomes they earn daily as a fish seller and a construction worker.

This importance of subsistence income to household livelihood strategies is not unique to the most marginalised sections of the precarious working class. Even for urban workers who have regular wage employment, social and kinship networks remain essential to long-term livelihood strategies. This is exemplified by another household from my fieldwork in Gauteng province. The primary respondent from this household is an unmarried man who works at a food-processing factory on the East Rand near Johannesburg. This man is a member of a union, but works in a subcontracted position. Like the rural household described above, unpaid household labour plays a major role in this man's long-term livelihood strategy. The man lives near his mother and aunt (who share a house), as well as his brother, who is a unionised metalworker. The mother and aunt own their house and plot. Since neither of the sons owns a house, the mother's home is central to their long-term livelihood strategies. The sons are using their savings to build a snack bar at their mother and aunt's house. Their plan is for the food worker son to eventually quit his job and run the snack bar full time. The other brother will join him when he reaches retirement age.

In contrast to the rural family from Mpumalanga, this household's long-term economic plan is much less developed. They have started to build the space that will be made into a snack bar, but it is not yet an established business and might not ever realistically be able to provide a living for the family. Yet it is still telling that, in describing his long-term economic strategy, this respondent focused on income generated with household labour rather than his wage work. He and his brother's energies and hopes are invested in his mother's home and his relationships to his family members. It is these kinship networks that both brothers see as the key factor in their long-term livelihood strategies. This exemplifies the importance of household-based livelihood strategies to South African workers’ own conceptions of their material interests.

Although it is difficult to make the leap directly from these practical understandings of livelihood strategies to workers’ political identities, it seems clear that there is a connection. An anecdote from a third interview subject illustrates the way in which precarious workers’ own identities often contrast with the ways in which labour scholars and unionists think about them. In rural Mpumalanga one interview respondent was a man in his 60s who lives in a house with his wife and grandson. The grandson, whose mother lives and works elsewhere, is a successful student in middle school and plans on seeking a scholarship to attend university. In the course of the interview the man was asked if he had a job and he said no. He answered that he took care of his grandchild and that he and his wife produced and sold small crafts like straw hats, ropes and grass mats. After a series of questions about household income, without uncovering any income sources beyond the crafts, which produced minimal profit, he was simply asked, where do you get money to live? After having earlier said he did not have a job, he now responded that he worked as a security guard at a local school. When asked how long he had been doing this, he answered 14 years! When I expressed surprise that he had said he didn't have a job despite working in the same place for 14 years, he explained that it was not a real job, just an extra source of income.

On the surface, it is puzzling that this man, who would be identified in the NIDS data as a precarious worker with regular wage employment, puts such a low priority on his long-held wage work that he doesn't even describe it as a job. Instead he identifies primarily as a husband, grandfather, craft maker and household head. However, when the long-term livelihood of the man and his household is considered, this identification makes more sense. If the school-age child he is raising were to grow up and attend university, the remittances he would be able to send would dwarf any increase in wages the man could earn by making demands on the rural school that employs him as a security guard. Raising this child, and maintaining a relationship with his employed mother, is this man and his household's best chance at achieving any level of long-term security. (Although that is not to suggest that he thought of his familial relationships so instrumentally.)

These households show the aspects of precarious workers’ social and economic lives that are not captured in survey data such as the NIDS. Particularly, they highlight the importance of subsistence work, which includes everyday unpaid domestic tasks but also extends to efforts to secure long-term livelihood. Like the NIDS data, these households illustrate the diversity of strategies that precarious workers employ to gain a livelihood. This diversity must be taken into account in debates about the way in which shifting experiences of work are shifting the politics of workers. In light of the findings presented here, the next section returns to the question that opened the article: what do the economic lives of South Africa's precarious workers reveal about a politics of precarity?

Conclusion: the emerging politics of precarity

It is difficult to decipher, in the South African case, a broadly shared unifying experience of precarious work. Precarious workers work in a wide variety of sites and under a diversity of conditions. What defines the economic lives of precarious workers is not the experience of work but the experience of insecurity. For a significant portion of precarious workers, the income they earn from their work is insufficient to provide for even the immediate needs of themselves and their families. For those who are able to gain a living through their work, the volatility of their income means that they cannot rely on their precarious work as a long-term livelihood strategy. As a result of this insecurity, precarious workers develop livelihood strategies that combine multiple sources of income. These livelihood strategies are usually centred on the household and draw heavily on the unpaid labour and social support of the workers and their kin. This broadly diversified household-centred livelihood is the central ‘common experience’ of precarious work. While the traditional politics of labour arises from the workplace – the key site of production – the politics of precarity arises from the home – the key site of reproduction.

This orientation towards the home shapes the way in which precarious workers formulate their identities and conceptualise their interests. Issues of work are neither unimportant nor irrelevant to precarious workers. Income from precarious work is valuable, but it cannot be relied upon in the long term. When workers think about their long-term strategies they tend to see their economic life away from work as the essential factor. In their identities they therefore tend to give priority to the roles that they hold within their households rather than their roles within the workplace or the labour market.

Of course, the household-oriented identities of precarious workers are not immutable. Social movements have, throughout history, often succeeded in changing the political identities of large portions of societies. The modern labour movement from the late nineteenth century through the middle of the twentieth century provides an example of one such success. However, the failure of contemporary labour organisations to incorporate precarious workers on a large scale over the past few decades demonstrates that building a political movement of precarious workers around issues of work faces serious obstacles. The data presented in this article suggest that if labour movement revitalisation is to be achieved, it will not be led by workplace-centred unions mobilising workers around the issues of traditional labour politics.

However, this does not mean that existing labour unions are inevitably destined to ‘run behind the new society’. The South African labour movement provides some innovative recent examples of existing unions which have begun to shift their attention towards the politics of precarity. South African unions have been at the forefront of struggles in the country around economic issues beyond the workplace, such as healthcare, transportation and land. The Congress of South African Trade Unions (COSATU), the country's leading labour federation, has been a key political supporter of the government's planned National Health Insurance (NHI) scheme, arguing that ‘the case for NHI is overwhelming’ (COSATU 2011). This support is noteworthy because, while some COSATU affiliate members (such as nurses and doctors) would benefit from increased expenditure on public health care, on the whole it is likely that the NHI would be a net transfer from COSATU members to the poorer majority of South Africans. Similarly, the union federation has been a leader in the fight against tolling of roads in Gauteng, the country's most populous province. The union has invested a great deal of its energy into this issue, which is not an issue of workers in the workplace but of individuals dealing with commodification in the realm of daily life. Another example of unions taking up pressing issues beyond the workplace comes from two of the country's more innovative unions, the Food and Allied Workers Union and the National Union of Metalworkers of South Africa (NUMSA). These two unions have launched an ongoing campaign about ‘agrarian transformation and land redistribution in South Africa’ (Numsa 2013). It is striking that NUMSA, which represents some of the country's most solidly industrial and urbanised wage workers, would take up the issue of land. Of course, the degree to which initiatives such as NUMSA's and the others mentioned have any meaningful organisational connection to precarious workers has been widely questioned. However, these shifts from ‘shop floor’ to ‘kitchen table’ issues among existing trade unions seem to offer confirmation that there is an ongoing shift from a traditional politics of a labour to an emergent, livelihood-centred politics of precarity.

No potential conflict of interest was reported by the author.