Gender and informal livelihoods Coping strategies and perceptions of waste pickers in Sub-Saharan Africa and Latin America

Purpose – This paper explores the perceptions and experiences of women and men who work as informal waste collectors in four different cities. The purpose of this paper is to map out how and to what extent occupational, political-legal, economic and social dynamics are experienced differently by gender in a highly vulnerable segment of the urban informal economy, and explore gender differences in these workers’ coping strategies and the levels of action they develop to protect their livelihoods. Design/methodology/approach – The analysis is based on a mixed methods study which combined a quantitative survey of informal workers with a qualitative participatory methodology. Study participants were drawn from a purposive sample of informal workers who belong to, or are affiliated with, membershipbased workers’ organisations. The sample consists of waste pickers (n1⁄4 614) from Belo Horizonte, Brazil; Bogotá, Colombia; Durban, South Africa; and Nakuru, Kenya. Findings – The data show that despite significant differences between women and men upon entry into (informal) employment, their perceptions of key drivers and impacts are largely similar, with the exception of concerns around various types of physical security among women. They also indicate that levels of action among men and women waste pickers are only moderately influenced by gender, but are strongly influenced by the degree of organisation in the sector and the symbolic assets held by workers. The findings also illustrate the way in which gendered power dynamics operate within the informal recycling sector and how different levels of sector organisation and development often contribute to opportunities for collective action and, in turn, a reduction in gendered vulnerabilities. Originality/value – The study offers a new policy angle which connects the level of sector organisation and development with the levels of action taken by informal workers in adapting to different types of shocks, as well as what this means in terms of gender empowerment.


Introduction
In most Latin American and Sub-Saharan African countries, a larger share of women than men are in informal employment (ILO, 2013b). Despite the greater share of the female workforce in informal employment in these two regions, gender inequalities in this sector are rife. Indeed, many of the disadvantages that women face within the informal economy in terms of earnings, working conditions, safety and security are now well-documented in the literature. Analyses of the source of these gender inequalities commonly focus on specific social constraints impacting women's entrepreneurship potential, including, though not limited to: gender differences in access to education, patriarchal and cultural practices that limit women's mobility, an unequal division of unpaid domestic labour, access to different social networks, greater exposure to occupational health and safety risks and less access to credit and start-up capital (Aterido et al., 2011;Folbre, 2006;Kabeer, 2003;Mahmud et al., 2012;Mayoux, 2001;Razavi, 2007;Wrigley-Asante, 2013).
While these factors may help predict modes of entry into livelihood activities, they do not account for gendered differences in perceptions of the key drivers of working conditions, nor are they necessarily strong predictors of livelihood-related coping strategies. Moreover, because many studies (see Dodlova et al., 2014;Grimm et al., 2012) focus on entrepreneurship as an outcome, little is known about the intersection of gender and livelihoods in segments of the informal economy where there is limited or less scope for extending lines of credit or promoting entrepreneurship. This study shifts the focus towards one such segment, the informal recycling sector, and analyses men and women's perceptions of their key livelihood challenges and the strategies they use to overcome them. This paper examines the perceptions and experiences of women and men who work as informal waste collectors or "pickers" in four cities in Sub-Saharan Africa and Latin America. The paper is motivated by two objectives. The first is to map out how and to what extent occupational, political-legal, economic and social factors are experienced differently by gender in a highly vulnerable segment of the urban informal economy. Second, the paper explores the coping strategies and, more importantly, the levels of action developed by workers to deal with the impacts from the factors listed above. The findings illustrate how gendered power dynamics operate within the sector, and how different levels of sector organisation and development contribute both to gendered vulnerabilities and to opportunities for collective action.
The paper is organised as follows. The first section reviews the literature on gender inequalities and constraints in the informal economy. Against this backdrop, an overview of the conceptual framework adopted for analysing vulnerabilities and strategies is presented. The following section outlines the data and methods used for the analysis. The third section presents the findings in two parts. In the first part, summary statistics detailing some of the key structural differences between men and women in the waste sector in terms of household organisation, income sources and earnings are identified. The second part examines the gender differences regarding enterprise characteristics and workers' perceptions of key challenges or "drivers" in the waste sector, their impacts, and coping strategies. Reflecting on how workers adapt to a variety of drivers, the analysis considers a broad spectrum of levels of action, ranging from the individual and household levels to the community level to city-wide collective action. The concluding section discusses the implications of these findings for other vulnerable groups of urban informal workers.

Informal waste collection
Informal waste picking is one of the most diverse yet least understood types of informal employment. Waste pickers are identified as those who undertake the primary collection or sorting of waste materials which have been "cast aside" (Dias, 2016;ILO, 2013a, b; 436 IJSSP 37,7/8 Samson, 2009a, b). These workers typically collect unwanted materials directly from households, various types of public places (e.g. streets, waterways or public parks), industrial or commercial buildings, or municipal landfills. A number of waste pickers, and particularly those who have formed workers cooperatives, even work in warehouses as waste sorters.
More broadly, however, relatively little is known about informal waste picking since national statistics (i.e. through labour force surveys) are often not effective at identifying and measuring employment in this sector. In the handful of countries for which at least some data are available, the contribution of the sector to total urban employment is less than 1 per cent (ILO, 2013b). The limited data available to measure the extent of waste picking in various countries also suggest that there is no clear gender dimension to the sector (ILO, 2013b). In other words, the gender composition seems to differ from one context to another and it is not clear whether this is due to data constraints or to the importance of local context in determining whether waste picking is more concentrated among women or men.

Gender inequalities and constraints in the informal economy
Causal theories on the informal economy from the past few decades have focussed on explaining why individuals enter informal employment as opposed to formal employment. The dominant schools of thought identify four main causes of entry into informal employment: exclusion of the urban poor from formal employment due to structural barriers; exploitation on the part of formal firms that create informal employment arrangements in order to avoid costs; voluntary choice or exit from formal employment by enterprise operators who seek to avoid state regulations; and strategic choice (Chen, 2012). Though each of these may explain the entry of some workers and entrepreneurs into informal employment, they focus little on gender disparities once in informal employment.
A more recent body of research examines the gender-related constraints on women's labour force participation (Abramo and Valenzuela, 2005), female entrepreneurship (Bushell, 2008) and women's economic empowerment (Mahmud et al., 2012). This literature identifies several policy-relevant individual-and enterprise-level constraints including access to formal education, skills training, credit and markets. Yet beyond these factors are more systemic challenges around translating women's economic empowerment to more far-reaching transformation towards just and inclusive cities (Moser, 2016). Moreover, while the bias in the economic empowerment literature is towards enterprise operators, many informal workers are sub-contracted, unpaid family contributors, members of cooperatives or work in more ambiguous employment arrangements that mask relations of power and subordination.
The waste sector is one segment of the informal employment pyramid where constraints on individual entrepreneurship such as access to credit are less relevant, and where the social dynamics of power more clearly intersect with gender dynamics. Examining the constraints and inequalities in waste picking through a gender perspective highlights the varying forms and degrees to which waste pickers experience exploitation and marginalisation. Some of these constraints may also be underpinned by the naturalization of hierarchical gender roles that also cut across status-based identities such as race and class.
Studies highlighting gender inequalities in the waste-picking sector reveal at least three forms of oppression women waste pickers face impacting their earnings, physical dignity and opportunities for political participation. On one level and related to working conditions, women waste pickers might not be allowed access to recyclables with the highest values, often receive less for the same recyclables collected by men, and face greater health risks as a result of handling waste and/or of working in insalubrious environments (Dias and Fernandez, 2013;Furedy, 1990;Muller and Scheinberg, 2003). On another level, women also do not find opportunities for occupying positions of authority within their work groups or 437 Gender and informal livelihoods when they do occupy those positions they may not be as respected as their male counterparts. As Muller and Scheinberg (2003) note, women's participation seems to be higher at the community and local level, with stronger gender gaps appearing as the distance from the community and the formality of the setting increases.
Broadly speaking, gender dynamics in the waste-picking sector refer to the ways in which the sexual division of labour manifests itself; gender-based violence emerges in relations with waste pickers, intermediaries and/or authorities; occupational and safety hazards impact men and women differently given specific biological needs; and gender stereotypes are impediments towards the economic and political empowerment of women workers. These expressions of marginalisation not only involve matters of distributive justice, but also entail the "deprivation of cultural, practical, and institutionalized conditions for exercising capacities in a context of recognition and interaction" (Young, 1990, p. 55). The following analysis builds on this emergent literature on gender and waste by examining men and women waste pickers' perceptions of key drivers in the sector and their strategies for overcoming the most serious challenges.

Conceptual framework
The analysis of livelihood strategies in this paper is disaggregated into two general categories: self-provisioning and claim-making (see Kabeer et al., 2013). The former refers to more individual-based strategies or problem solving among one or two waste pickers. A second feature of self-provisioning strategies is that it is physically restricted to the waste pickers' work environments. In contrast, group claim-making strategies involve a broader scope of interaction, including demands on relatives, neighbours, community, NGOs or different levels of government (Chambers and Conway, 1992). More recent debates on claim-making regarding gender equality have emphasised either institutional responsiveness to claims (Htun and Weldon, 2010) or how claim-making by women workers can be aligned with a politics of recognition, redistribution and representation (Kabeer, 2015). The paper seeks to contribute with a reflection on the processes leading to claim-making prior to the ability to more formally direct claims at institutions or to have them supported by local and global actors as discussed in the work of Kabeer (2015). Hence, the analysis of claim-making strategies here considers the ways in which waste pickers engage with or have access to different actors[1] in order to minimise the threats to their livelihoods. This strategy also allows us to trace group claim-making based on the formation of a collective identity associated with level of sector development and organisation.
Due to the level of vulnerability experienced by informal workers, and specifically by women waste pickers, the idea of claim-making here reflects an initial process of relying on collective action as a coping strategy. In other words, it includes making demands on informal and formal institutions or groups, not necessarily just directing demands at different levels of the state. In this sense, collective action should be understood as a fundamental symbolic asset that paves the path for making claims on more formal and institutionalized levels. In other words, it can promote workers' better understanding of how claim-making cannot be disassociated from economic, social and political injustices. As a symbolic and intangible asset, collective action fosters a stronger worker identity and self-dignity (Hill, 2001) that activates voice and agency.
The analysis in this paper also rests on a structured comparison of waste-picking sectors in four cities, two Latin American and two African. The participatory component of the research offers the advantage of "hearing local people's voices and priorities", while the application of the same data collection tools in cities from two regions draws out important variations in the regional context that impact specific livelihood conditions and strategies (Moser and Stein, 2011). In particular, 438 IJSSP 37,7/8 the paper introduces the idea of the "level of sector development", meaning the extent to which the value chain in recycling has consolidated around a set of major, organised actors. In the Latin American cities, the legal framework, waste cooperatives, intermediaries and formal sector players are well defined, whereas in the African cities all of those remain in early stages of formation.

Study design and data sources
The data analysed in the paper were collected as part of a ten-city study, the Informal Economy Monitoring Study (IEMS), which investigated the working conditions of urban informal workers on three continents and in three occupational sectors (street vending, home-based work and waste picking). The ten cities were chosen through their participation in an ongoing collaboration between organisations of the working poor, international alliances of membership-based workers' organisations (MBOs) and support organisations that are concerned with improving the livelihoods of informal workers. Given the well-documented challenges with collecting representative data on informal workers, the IEMS was designed to conduct detailed research on a small sample of workers over time.
In order to collect this longitudinal data, it was important to identify informal workers from a database which contained a priori information on the key demographic and work-related characteristics of the workers' organisations. The data analysed in this paper come from the baseline wave of the IEMS (in 2012) [2].
The IEMS is a mixed methods study which combines a quantitative survey of informal workers with a qualitative urban participatory methodology. The fieldwork was conducted by local researchers in each city after attending an intensive, week-long training course in South Africa. The survey questionnaire was piloted extensively and then translated into a number of local languages. The surveys were conducted with workers on site with the assistance of handheld electronic devices or "personal digital assistants". Once each interview was completed and logged, the data were uploaded to a central database which was managed by the research team. As a result, the survey data from all ten cities were processed, captured and checked for quality in real time. About half (n ¼ 75) of the survey respondents in each city were then selected to take part in a series of focus groups. The design of the focus groups was adapted from a participatory methodology developed by McIlwaine (1999, 2004). The focus groups consisted of a number of ranking, listing and mapping exercises which have been shown to be useful in collecting data on urban livelihoods (see Moser and Stein, 2011). Building on previous experiences with this methodology, the compositions of the focus groups were designed to promote different dynamics ( e.g. the research team organised single-sex as well as mixed groups) and, at all times, the focus group facilitator aimed to encourage "visual rather than written or verbal accounts of situations or issues" (Moser and Stein, 2011, p. 468).
Study participants were drawn from a purposive sample of informal workers who belong to MBOs. As such, the data are not representative of all workers in these sectors or cities[3], though they are designed to be as representative as possible of workers who are MBO members or affiliates. Detailed information on the characteristics of each organisation's members was made available through the organisations themselves. Based on the administrative records of the organisations, we selected the participants based on a purposive quota design in order to ensure that the characteristics of our sample resembled the characteristics of the organisations in each respective city. Therefore, while not representative, the sample reflects the memberships of these organisations in terms of key characteristics such as gender, type of activity and location of work. Consultations with members of the organisations helped the research team identify the criteria with which to stratify the sample in each city. The results presented in this paper are drawn from a sample of waste pickers (n ¼ 614) from Belo Horizonte, Brazil; Bogotá, Colombia;

439
Gender and informal livelihoods Durban, South Africa; and Nakuru, Kenya. The criteria for the design of the quota sample in all four cities were gender and place of work (e.g. a fixed source of waste vs itinerant collectors) (Table I).

Context
The characteristics of the waste-picking sector across the four cities in which fieldwork was conducted are vastly different. Perhaps most importantly, waste pickers are considerably more organised in Latin American cities (and particularly in Brazilsee ILO, 2013a;Dias, 2011) than in other regions. Indeed, the respondents from Belo Horizonte and Bogotá were largely members or affiliates of workers cooperatives or associations which are well established in the local recycling value chains. The data presented in Table II show that nearly all study participants from Belo Horizonte and just under a quarter from Bogotá are the organised members of cooperatives (though all are affiliated with workers' associations). At the same time, the waste-picking sectors in the two African cities are relatively underdeveloped and the workers in these cities are only beginning to organise into associations or collectives.

Household structure, income sources and earnings
The data presented in Table III show that, as expected, women live in households which are structurally different from the households in which men live. In particular, female waste pickers live in larger households, live with a greater number of children and have far higher dependency ratios[4] than male waste pickers. This is particularly the case for the sample from the two African cities where women live in households with a dependency ratio two times larger than the households of their male counterparts (0.76 and 0.35, respectively).
Since household structures differ between women and men, it is not surprising that there are differences in household employment characteristics and in access to the labour market  Table IV ). While, on average, the women waste pickers who participated in the IEMS live in households with a larger number of workers (1.85 in the African cities and 2.28 in Belo Horizonte and Bogotá) relative to their male counterparts, the employment ratios in the households of female waste pickers are lower than in the households of male waste pickers (largely due to higher dependency ratios). This means that, even though many women live with other informal workers (particularly in the African cities), their earnings are spread more thinly over a larger number of household members, and particularly children.
The data suggest several interesting differences in the way that women's earnings from waste picking contribute to the income of their households (Table IV ). In the two African cities, women's earnings are significantly less likely to be the main source of household income (58 per cent of women and 82 per cent of men from the sample provided the single largest source of income to their household). When female waste pickers are not the primary breadwinners, the earnings of another informal worker in the household is often the main income source in the African sample (i.e. 17 per cent of women). In the two Latin American cities, on the other hand, the earnings of women are only marginally less likely to be the household's main income source. In fact, the structure of household income and the contribution of women's and men's earnings to their households are very similar in the Belo Horizonte and Bogotá sample.
Table V now looks specifically at the earnings from waste-picking activities, by gender, in the two regions. The main finding is that gender differences in earnings are considerably larger in the two African cities. Both mean and median earnings are far higher for male waste pickers in Durban and Nakuru (e.g. median monthly earnings for women are USD45 and USD70 for men). However, the most appropriate comparison category is median hourly earnings (since men work considerably more hours than women and because the earnings distribution is somewhat skewed). After adjusting for hours worked, men still earn more

441
Gender and informal livelihoods than women in both regions (e.g. USD2/hour for women and USD2.31 for men in the Latin American cities), confirming the income gender gap identified in the broader literature on informal employment. The last row of the table shows that female earnings from waste picking are about 71 per cent of male earnings (controlling for hours worked) in the African cities and 89 per cent in the Latin American cities. Men therefore earn more than women in our sample but the differences are considerably largely in the two African cities.
One of the explanations (Glick and Sahn, 1997;Kucera and Xenogiani, 2009) for gender inequalities in informal employment from the broader literature is that women, in this sector, have far lower levels of education relative to men. The IEMS data provide clear evidence ( Table VI) that in both regions, there is a large gender disadvantage in access to education. In the two Latin American cities women are twice as likely as men to report not having any schooling or education at all (12 per cent and 6 per cent, respectively). In Durban and Nakuru, men are also significantly more likely to have attained at least some secondary education or to have completed secondary schooling.

Assessing perceptions on vulnerabilities
Against the backdrop of some of the expected (and measureable) gender differences outlined in the previous section, assessing the qualitative perceptions of waste pickers reveals the ways in which certain drivers sustain gender vulnerabilities. The findings identify some of these vulnerabilities by tracing the workers' exposure to shocks and risks and explores how they experience occupational, political-legal, economic and social factors differently. Despite the fact that vulnerabilities, in general, create a sense of powerlessness, the central argument here is that specific forms of vulnerability can undermine a workers' sense of dignity, limiting the levels of actions taken. From a

442
IJSSP 37,7/8 gender perspective, these vulnerabilities establish further barriers for women to expand their coping strategies beyond an individual level.
The analysis in this section draws from qualitative data from 40 focus groups which were conducted across the four cities [5]. During the focus groups, workers were asked to engage in a listing and ranking exercise to identify the factors that most hinder their work and livelihoods. Notwithstanding the gender differences presented in the first section of the findings, the focus groups did not reveal any clear gender or regional contrasts with regard to the perceptions of negative drivers. The perceptions of both women and men highlight how economic and occupational drivers are the forces that primarily underlie social exclusion (Table VII).
Value chain drivers appear as the predominant negative driver in both regions. The majority of the participants referred to low selling prices as the main problem linked to value chain dynamics, though they also mentioned being cheated by middlemen, the scarcity of materials and competition from other waste pickers. Waste pickers also discussed precarious working conditions, such as the lack of proper infrastructure and/or storage facilities and health and safety risks, as the second most important negative driver.
Another commonality across the cities is related to the ways in which city and government practices create challenges for workers, including but not limited to the confiscation of materials. The participants in Nakuru, Durban and Bogotá made more references to verbal and physical harassment from local officials. As a male respondent from Bogotá explained: "The district government wants to take the recyclable materials from us; they deny us the right to work; they want to get rid of us". In Durban, South Africa, a female waste picker similarly identified how such harassment prevents her from accessing resources. "Police bring their dogs when we are getting the goods from the cars and they chase us away with these dogs" (Durban FG 1). The perception was more nuanced in Belo Horizonte, where the participants viewed their relations with city officials as both negative and positive. Of the four cities, Belo Horizonte has the longest history in terms of efforts to include waste pickers in its municipal solid waste system, dating back to the 1990s (Dias, 2011). Admittedly, workers in this city still expressed concerns about threats made by the local authorities to incinerate waste before it can be sorted or the lack of opportunities available for strengthening ties with the municipality.
Both men and women across the cities emphasised the discrimination experienced in their interactions with the community. As one male waste picker claimed in Durban: "People undermine us because of the way we make our living, picking waste" (Durban FG 10).

Gender and informal livelihoods
The participants in all four cities repeatedly claimed they are mistaken as thieves or homeless persons. A female waste picker in Nakuru expressed her concern by explaining that: "We cannot be differentiated [by the community] from criminals who occasionally sneak to the dumpsite" (Nakuru FG 1). Others claimed that "some people look at the homeless people and then look at us, and they get the same image" (Bogotá, FG 7).
Men made almost double the amount of references related to the problems with the community in comparison to women in both regions. For example, in Durban and Belo Horizonte it was noted that men are more commonly itinerant collectors[6] and women are either working at fixed collection sites or in sorting warehouses. In this sense, given that men are circulating in public spaces, they may be facing harassment from both local authorities and the community more so than women. Despite the fact there are relatively few cultural restrictions on the women's mobility in both regions (Kabeer et al., 2013), this difference may point to a gendered division of labour within the sector, where the activities designated to men and women mirror the public-private divide and reinforce stereotypical gender roles.

Impacts of drivers and gender-specific vulnerabilities
During the focus groups, the participants were asked to identify the two most important negative drivers through a listing and ranking exercise. In a follow-up exercise, the participants were then probed to discuss the causes and impacts of these drivers. Table VIII classifies the answers into three broader categories of impacts: earnings, productivity and dignity.
References to earnings include impacts on physical assets, a decrease in income and consequent reduction in household consumption and difficulty in gaining access to materials. Impacts on productivity refer to the responses highlighting the ability to work efficiently, such as exclusionary laws in the city that lead to harassment and arrests from authorities and precarious workplace conditions. Impacts on personal security and dignity refer to mentions of threats to workers' well-being. The data in Table VIII therefore illustrate the workers' perceptions of how they are impacted by the negative drivers identified in Table VII or, in other words, how workers perceive the connections between drivers and impacts. With the exception of women in Sub-Saharan Africa, waste pickers from both regions made more references to impacts on earnings and productivity. These references reveal how workers understand their status in the value chain hierarchy and lack of proper infrastructure for handling waste are directly linked to these impacts.
While workers generally referred to the impacts in a similar fashion in all cities, the qualitative findings indicate an observable gender difference with regard to personal security and dignity in Nakuru and Durban. These references to personal security and dignity can be divided into three broader dimensions: occupational health and safety, harassment and violence and dignity and respect. Occupational health and safety is the most referenced problem, appearing in 19 of the 40 focus groups. In addition, these problems were mentioned almost three times more by women in the African Saharan cities than in the Latin American ones. Broadly speaking, the impacts on personal security and dignity reinforce the assumptions that different forms of gender-based vulnerabilities are exacerbated in the contexts of inadequate basic infrastructure and access to services (Tacoli, 2012).

Gender-based biological needs and violence
In all of the focus groups, both women and men highlighted a number of occupational risks they encounter. However, when women made references to occupational health and safety, there was often a clear relation to their specific biological needs. In Durban, for example, there were several mentions to problems related to the lack of sanitary toilets. In several instances, these discussions also evidenced women's fear of sexual violence. Out of the seven individual mentions on the impacts of not having toilets, three were directly emphasising the fear of being raped. Women also mentioned the consequent health impacts of such workplace conditions: "We are afraid that this is going to cause us to get a lot of diseases, as we are already getting rashes because of these men when they are spraying the dust it does affect us" (Durban FG 1).
While the qualitative data show that workers are susceptible to such forms of harassment, it also suggests women, particularly in the African cities, are dealing with institutionalized and economic forms of gender-based harassment and violence in their interactions with city officials, middlemen and male waste pickers. In Nakuru and Durban, women's focus groups revealed how gendered power relations among workers are played out along the lines of "territorial or gender-based 'turf' violence" (see also Moser, 2004;Moser and McIlwaine, 2006). In addition to fear of thieves, women mentioned being intimated by male waste pickers. As a woman noted in Nakuru: "When the dumpsite is full, the vehicles don't bring waste. Sometimes when those vehicles come, the boys/men refuse women from getting on the vehicles to get the materials, so the boys get the best of the materials" (Nakuru FG 1). Similarly, in Durban, a female participant claimed that "the boys just push us and they go and get better things than us" (Durban FG1).
These narratives provide further insights into some of the reasons that male waste pickers' mean and median earnings are higher in both Durban and Nakuru. Thus gender here serves as a way of establishing hierarchies within the value chain, revealing a clear economic and social dimension to personal security. Part of the reason for this may be that, in contrast to Belo Horizonte and Bogotá where workers are organised to different degrees, waste pickers in Nakuru and Durban seem to be operating on a more individual level. The fragility of women's livelihoods in these cities is intensified in light of physical competition with men, which may further diminish the possibilities for collective action or result in women engaging in less profitable work (see also Chant and Pedwell, 2008).
How women and men cope in the waste sector Turning now to an analysis of strategies for coping, Table IX shows that men and women in both regions are overwhelmingly resorting to self-provisioning strategies. Many of the

445
Gender and informal livelihoods self-provisioning strategies result from attempts at individually mitigating the impacts of value chain forces, workplace conditions and city or government policies. The data also highlight how women in the sample from the two African cities are almost four times less likely to resort to claim-making strategies when compared to men in both samples and women in the two Latin American cities. Ultimately, women in Nakuru and Durban cities are mainly adapting to or enduring hardships.
The qualitative data (Table X) also illustrate the most frequently mentioned types of self-provisioning strategies. Among these strategies, the workers claimed they diversify their work activity, adapt to their work environment, endure hardships and reduce household expenditures. In the Latin American focus groups, some self-provisioning strategies included joining with other workers to strengthen their organisations. These strategies were categorised as self-provisioning strategies because they were utilised prior to the ability of the workers to make a demand based on a collective identity. In these cases, workers handled a problem within the confines of the workplace.
When referring to adapting to hardships as a strategy for dealing primarily with workplace conditions and value chain forces, the workers reported that they wear protective equipment, take over-the-counter medication for pains or bandage wounds themselves or even choose not to work for a short period of time. In situations involving some form of physical threat, particularly in Nakuru and Durban, women claimed that they hide from authorities or find a safer place to store materials. The differences being reported by men are that their strategies for adapting to circumstances seem to work towards diminishing the economic impact of problems. For example, men in Bogotá try to "dress up" and "wear identity cards" when picking up materials in order to avoid harassment. These examples point to the importance of how belonging to an organisation fosters a sense of a professional identity, which is a basis for workers to make claims.
Another gender difference noted in terms of self-provisioning strategies relates to the option of diversifying activities to supplement income. Both men and women waste pickers looked for stereotypically "masculine" or "feminine" jobs, which in the case of the latter, reinforces women's roles as caretakers. For example, men in Nakuru find jobs working at construction sites, mending or repairing shoes or cleaning drains, while women wash clothes and plait hair. Similarly, men in Belo Horizonte commented on the fact they pick up debris at construction sites, whereas women often mentioned they try to look for a cleaning services job. Furthermore, both men and women mentioned reducing expenditures within the household as a coping strategy but there were far more mentions among women in Latin America. All-women focus groups mentioned cutting back on expenses 20 times with half referring to food rationing. Men in these cities only mentioned reducing household budgets four times.
More than comparing the number of individual mentions to claims being made on distinct levels, the findings above help draw out how the level of sector development and  organisation are linked to workers' claim-making strategies. Overall, there are stark regional differences within the sample. The group of waste pickers from the two Latin American cities reported making almost double the amount of claims (51 mentions) compared with waste pickers in the African sample (28 mentions). Furthermore, 78 per cent of the claims in Nakuru and Durban are made at the local, community level, while in Belo Horizonte and Bogotá the claims are almost evenly split between the community and political levels (49 and 45 per cent, respectively).
Another facet to consider is how regional differences are even more pronounced when considering women's coping strategies in our sample. Women in Belo Horizonte and Bogotá were five times more likely to make demands than women in Nakuru and Durban. Nevertheless, both men and women reported making more claims in Bogotá than in Belo Horizonte, which may be attributed to the strengthening of dialogue in the former between workers and different levels and branches of government. Participants in Belo Horizonte, on the other hand, expressed greater dissatisfaction with the current possibilities for dialoguing with the local government than in the past (Dias and Samson, 2016). Thus, the ability to make claims, particularly at the political level, may be associated with a favourable political environment or with areas where the state is active and can deliver or meet the demands (see Houtzager et al., 2002).
It is also important to consider what kinds of demands are being made at each level in order to establish connections between the type of claim and sector organisation. In the African cities, most of the demands made by women ( five mentions) and men (ten mentions) at the community level are related to seeking some kind of medical assistance due to precarious work conditions. As a participant from Durban declared: "We have to go to the clinics when we are sick, at the clinic they will refer you to the hospital" (Durban FG 1). However, men reported attempts to mobilise the community or building administrations to facilitate the process of collecting materials (three mentions), indicating they may have stronger social networks to leverage.
In a similar fashion, community-level demands in Bogotá indicated that workers prioritise their relations with the community in almost an entrepreneurial fashion. "We learn to be friendlier with the residents to persuade them to leave material for us" (Bogota FG 9), claimed a participant. Another worker stated this is a part of the process of "making yourself known in the sectors so that people will set aside material for you, and they even sometimes give you a bit of money. It's marketing, if they have a good opinion of you, they will call you" (Bogotá FG 8).
In total, 83 per cent of the demands at the community level in Bogotá referred to negotiations with building administrators to gain access to recyclables. As one participant noted: "[we] speak to the administrators of fixed sources, because some administrators allow you to collect if they see that you are organized and there is an organization that can answer for us if need be" (Bogotá FG 2). Belonging to an organisation creates a shared identity that helps these workers build on their financial and social assets.
In addition to the aforementioned assets, organisations can also work towards changing society's negative perception of workers. What should be emphasised here is the possibility for diminishing social stigma and effectively building workers' symbolic assets. As one participant explained: "We must participate actively with [our MBO]. They help us expand our sources by using photographs and videos to develop a good image for the organization" (Bogotá FG 10).
The establishment of a worker identity is fundamental for claim-making at the political level as well. In Latin American cities where women belong to MBOs, the demands made at political level include participating in marches, protests, signing petitions and negotiating with key political actors. The scope of women's actions in Bogotá, for example, include "open[ing] dialogue with police" (Bogota FG 4), "speak[ing] with the local mayor to ask for permission to work" (Bogota FG 4), as well as being represented at "all levels of government" (Bogotá FG 5).

Gender and informal livelihoods
Such demands are nonetheless tied to another role the organisations play in terms of disseminating information and providing capacity building courses. Knowledge on workers' rights may feed into the strategies for making claims on both a political and economic level. In these cases, workers in Bogotá were more prone to negotiating with middlemen to "establish the price" (Bogotá FG 6) or understand "why they lower the prices" (Bogotá FG 3). Nevertheless, none of the waste pickers across the four cities mentioned making claims to formal financial institutions. This reinforces the difficulties these informal workers have in gaining access to credit.

Conclusions and implications
This paper has explored gender and livelihoods across four different cities through the lens of a particularly vulnerable segment of the urban informal economy. Its contribution to the literature is twofold: first in identifying gender similarities and differences in entry, perceptions of key drivers and coping strategies; and second in identifying the regional differences that intersect with gender gaps through a structured comparison of qualitative and quantitative data.
In terms of the first contribution, by focussing specifically on the workers' perceptions of the key drivers of their livelihoods and work places, it has identified several important differences in the way in which women and men respond to negative drivers. This is particularly important since the analysis has shown that, in line with the existing literature, a number of well-documented sources of gendered vulnerability are present in the informal recycling sector. However, the findings also show that, despite the structural differences in the way that women and men enter into informal work, their perceptions of the factors which influence their livelihoods, both positively and negatively, are remarkably similar.
Yet in terms of the second contribution, the regional comparison shows that in cities where value chains are less developed and workers less extensively organised, gender differences regarding safety and dignity are far more pervasive and levels of action more confined to the individual and household levels. The study thus presents a new policy angle which connects the level of sector organisation and development with the levels of action taken by waste pickers in adapting to different types of shocks, as well as what this means in terms of gender empowerment. In less developed sectors, women are not only more subject to physical safety concerns, but also to institutions and actors that not only erode tangible assets, but intangible ones as well.
The analysis also suggests that, in less developed sectors, women in particular are not only denied recognition of the work they do, but also struggle to protect themselves from different forms of harassment and violence, as well as particular health and safety issues. Their even more marginalised position within the value chain may reinforce patriarchal norms and gendered power relations that naturalise their subordinated status. This is in contrast to what happens in more developed sectors where some level of collective identity has been established, even if not on a gender basis.
In cities with more developed sectors, women waste pickers seem to be coping through a broader range of levels of action because they envisage room for making negotiations and claims either on a community level or political one. In these cases, their symbolic resources serve as the catalysts for making claims to remedy the numerous economic and political injustices they face. This is also associated with the fact that they can rely on the support of their MBOs and NGOs that encourage collective action and help strengthen their demands through institutional channels.
Put differently, the qualitative findings provide insights as to how organising workers can provide opportunities for accumulating both tangible and intangible assets. This is particularly true for women who tend to limit their coping strategies to an individual level. Since many of these women are located at the most invisible margins of the informal 448 IJSSP 37,7/8 economy and cultural practices have undervalued women's productive and reproductive labour (Kabeer et al., 2013), it is only natural that many are not compelled to make demands in a more collective manner. This became clear in cities such as Nakuru and Durban where women were finding ways to endure hardships on various levels.
On the other hand, organising workers seems part of the path towards diminishing these negative impacts on livelihoods and forging demands on a collective level. In this case, symbolic assets are important because they are foundations for workers, especially women, to envision other ways of protecting their livelihoods, including through the expansion of social networks and negotiating channels. While the findings do not diminish the importance of building upon other assets, they shift attention to how symbolic assets are important in light of the gender vulnerabilities related to physical security and dignity women waste pickers experience. This particular conclusion resonates fairly closely with what Kabeer et al. (2013, p. 260) identify as the "politics of recognition" where the most marginalised of workers are concerned as "much about dignity as daily bread".
In sum, from a policy perspective the key implications of these findings are that access to credit and other financial barriers are far less important in some occupational sectors than others, and that value chain integrations and basic protections of physical safety and human dignity would have a more transformative impact on livelihoods in the most vulnerable segments of the informal economy than financial sector interventions. Such protections would not only benefit these workers, but also the households that depend on their earnings.

Notes
1. For studies that use a similar categorisation to understand the efforts of local popular organisations in making demands, see Houtzager et al. (2002). Our objective in this paper is to capture any strategies that point in the direction of more group-based actions and to understand what may be the potential factors that influence this type of coping strategy. The paper will not cover the ways in which organising workers influences the process of making demands on the state and the levels of government at which the demands are directed.

2.
A follow-up round of data collection is planned for 2017.
3. Detailed and representative data on the key characteristics (e.g. gender, status in employment, location, main working activities) of workers at the city-wide level are not available.
4. The household dependency ratio can be interpreted as a proxy ratio of household members typically not in the labour force (dependents, e.g. children) to those typically in the labour force ("productive" householders). A higher household dependency ratio therefore denotes a larger number of dependents being supported by each economically active household member.
5. Each city team conducted a total of 15 focus groups: five with only women, five with only men and five with a mix of women and men. This analysis considers only those focus groups that are exclusively made up of women and those exclusively made up of menhence ten focus groups per city in each of four cities.
6. The survey data, for example, show that about 44 per cent of men and only about 20 per cent of women from the two Latin American samples collect waste from street sites (as opposed to fixed sources such as landfills).