‘But now I dream about my house’: women's empowerment and housing delivery in urban KwaZulu-Natal, South Africa

Abstract

Since 1994 the post-apartheid government has been investing in low-cost housing delivery. Besides providing funding, the state's housing delivery process has aimed to empower communities, particularly women. Using data collected in communities in urban KwaZulu-Natal through focus group discussions and household interviews, this paper examines the extent to which women have been empowered by their participation in housing delivery. The findings show empowerment has been achieved at different levels in the housing delivery process. The paper argues that to understand the extent to which it has been achieved it is important to examine all of the different aspects of the housing sector and the intensity of women's involvement.

Keywords:

• Women
• urban
• post-apartheid housing
• empowerment
• South Africa

1. INTRODUCTION

In the development literature, the notions of participation and empowerment are often discussed together, the assumption being that participation leads to empowerment. The idea of empowerment features in discussions about self-help, but to a lesser extent than participation. Set in the post-apartheid policy context, this paper critically examines the notion of women's empowerment. It first looks at the extent of women's involvement in housing delivery, drawing on empirical data to highlight their perceptions of how the process has empowered them. It then develops a framework for locating women in housing and establishing their level of empowerment, again drawing on empirical data. The paper concludes by summarising the key issues arising from the study and drawing out the implications for women's empowerment.

2. THE POLICY CONTEXT OF WOMEN'S EMPOWERMENT IN SOUTH AFRICA

The South African Constitution states that ‘everyone has the right to adequate housing’ (section 26 (1)), and goes further to underscore that ‘the state must take reasonable legislative and other measures, within its available resources, to achieve the progressive realisation of this right’ (Republic of South Africa [RSA], 1996:section 26 (2)). The right to adequate housing is embedded in the Bill of Rights. Against the background of the post-apartheid housing policy, the present paper notes that since 1994 the democratic state has taken an enabling approach to housing delivery, a notion encapsulated in the White Paper (RSA, 1994) and the Housing Act 107 of 1997. In the Housing Act (RSA, 1997), a guiding principle of housing development is that the state is expected to:

encourage and support individuals and communities, including, but not limited to, cooperatives, associations and other bodies which are community based, in their efforts to fulfil their own housing needs by assisting them in accessing land, services and technical assistance in a way that leads to the transfer of skills to and empowerment of the community. (RSA, 1997:part 1 (d))

In this clause is inscribed the notion of empowerment as an element of housing delivery.

Women's empowerment has been a key focus of the post-apartheid state, as can be seen in various policy documents. Affirmative action is one of the policies that require the representation of women and blacks who were previously disadvantaged by apartheid in all sectors of the economy (RSA, 1998, 2003). The concern with women in the housing sector came into focus in 1998 when the women's reference group was formed (National Department of Housing [NDoH], 1999) to advise the Department of Housing on gender issues. From 1998 it took another 8 years for the Department to concretise its ideas on gender, and only in 2006 did it devise a programme for mainstreaming gender in housing as a sector of the economy (NDoH, 2006). The document on gender mainstreaming calls for the disaggregation of data by gender, and points to processes for ensuring women's empowerment and achieving gender equality but fails to explain what empowerment entails in the context of housing. The post-apartheid housing policy in South Africa provides for state-aided self-help, where the government assists by offering a housing subsidy. The state's role here is to regulate the policy environment and provide finance. According to the White Paper, the aim is to enable ‘a development process driven from within communities … and empowering people to drive their own economic empowerment; the development of their physical environment and the satisfaction of basic needs’ (RSA, 1994:23). However, although women have been seen to participate in the housing process it is not clear to what extent their participation has contributed to their empowerment, either as individuals or as groups.

The Construction Sector Charter (RSA, 2006), on the other hand, recognises the low representation of blacks in the sector and sets targets to ensure their greater participation and ownership. Specifically, the Charter sets a target of 30 per cent black economic interest in the sector and 10 per cent economic interest by black women by the year 2013. The Department of Housing has so far provided 3 043 900 subsidies and built 2.4 million houses (NDoH, 2007); the data do not give a gender breakdown of the beneficiaries.

Organisations such as Women for Housing (WfH, 2007), formed in 1995, and South African Women in Construction (SAWIC), formed in 1997 (see Siphayi, 2008), were operating long before the document on mainstreaming gender (NDoH, 2006) was drafted. WfH draws membership from a range of housing professionals but focuses on offering training, advocacy and support to women in housing construction. SAWIC, on the other hand, draws its entire membership from women in the construction sector and seeks to empower women by intervening to help them access jobs, training, finance and networks in the construction industry. Like WfH, SAWIC has chapters in all nine provinces of South Africa. Although there are other organisations that generally focus on women's empowerment – such as the South African Women's Entrepreneurship Network – the WfH and the SAWIC are the two organisations that deal directly with empowerment of women in the housing sector.

3. THE TERM ‘EMPOWERMENT’

The term ‘empowerment’ can have many meanings and is used in a wide range of contexts (Rowlands, 1997; Kabeer, 1999; Townsend, 1999; Klasen, 2006; Schuler, 2006; Chant, 2006; Beteta, 2006). It is a term that has been embraced by conflicting schools of thought – neoliberals, neo-Marxists and Third World grassroots groups – to denote whatever the user wanted it to mean (Rowlands, 1997). Townsend aptly notes that:

the participation of local people in a project can mean involvement in all design and implementation, or merely a show of consultation which is never intended to be allowed to change anything … Similarly empowerment can be used to describe poor people finding the power to help themselves, or a government privatising health or pensions so that it can cut taxes in order to ‘empower’ citizens to decide whether to spend their income on health or pensions or conspicuous consumption. It now seems to refer to promoting entrepreneurial confidence among poor women rather than any need for the powerful to change, as if the only problem is the lack of drive among poor women. (1999:21)

Empowerment, according to Rowlands, is ‘A process whereby women become able to organise themselves to increase their own self-reliance, to assert their independent right to make choices and to control resources which will assist in challenging and eliminating their own subordination’ (1997:17). The term means more than just individual empowerment. It has political and economic dimensions; it includes gaining the confidence and ability to know and negotiate for rights, from the micro to the macro level – that is, from the private (household) to the public sphere of economics and politics. Participating in identifying needs is an essential part of the process of empowerment. The control of resources as the means to empowerment is also essential, although only part of a wider purpose.

Moser postulates that empowerment is:

the capacity to increase their own self-reliance and internal strength. This is identified as the right to determine choices in life and to influence the direction of change, through the ability to gain control over material and non-material resources. (1989:1815, quoted in Rowlands, 1997:17)

Like Moser (1989), Kabeer (1999) underscores the idea of choices in understanding the notion of empowerment:

the notion of empowerment is inescapably bound up with the condition of disempowerment and refers to the processes by which those who have been denied the ability to make choices acquire such an ability … empowerment is about change, it refers to the expansion of people's ability to make strategic life choices in a context where this ability was previously denied to them. (Kabeer, 1999:437)

Kabeer (1999) conceptualises the ability to make choices in terms of three dimensions – resources (preconditions), agency (process) and achievements – that are perceived to be interrelated and uses these to determine the extent to which women are empowered. Whereas Kabeer's focus is on choices that individuals make, Rowlands's (1997) conceptualisation sees participation as central to understanding empowerment.

Young suggests that, despite differences in how feminists conceptualise empowerment, in the ordinary sense:

empowerment is about people taking control over their own lives: gaining the ability to do things, to set their own agendas, to change events in a way previously lacking. This may include affecting the way other people act consciously or unconsciously forcing changes in their behaviour. (1997:371–2)

Feminist interpretations of empowerment move beyond formal and institutional definitions of power and incorporate the idea that the personal is political. Men's dominance and the pervasiveness of women's subordination point to a multidimensional approach in achieving their empowerment. According to Barroso (1997), the process of empowerment has to occur at several levels, with transformation leading to change in the ideological system, in access to resources and in institutions and structures at several levels, such as the family and the household, the village and the community, the state, the market and the international structures. Townsend suggests that ‘The goal of women's empowerment is not just to change hierarchical gender relations but to change all hierarchical relations in society, class, caste, race, ethnic, North-South relations’ (1999:19). Just as women have to work collectively to gain self-worth and understanding of empowerment issues, men too should reflect on the way their power is a double-edged sword, because it structures their relations with other men in competition and conflict, and makes cooperation highly problematic. Empowerment goes beyond participation in development to include transformation: it implies validating women's way of thinking, their intelligence, their experience and abilities.

In the South African context, various forms of social power and gender relations have to be addressed concurrently. It is impossible to address gender as a separate set of inequalities without addressing race and class, as these are intricately interwoven in women's lives. Empowerment is not only an external process but a process that has to bring about intrinsic changes not only in women but also in men. The above discussion suggests there is a need to establish the level of empowerment in housing delivery in South Africa.

Debates on how women's empowerment might be measured abound. One of the ways of measuring it is through what Kabeer refers to as ‘universally valued functionings, those which relate to the basic fundamentals of survival and well-being, regardless of context’ (1999:439). These include such things as adequate housing, food and health – that is, basic needs.

According to Kabeer, measurements of women's empowerment should focus on three dimensions: resources, agency and achievement. Resources are the human, material and social capital at the disposal of individuals, accessed through a range of social relationships such as the family, market and community. The rules and norms that govern these resources determine who has access and who is excluded. Agency is ‘the ability to define one's goals and act upon them. Agency is about more than observable action: it also encompasses the meaning, motivation and purpose which individuals bring to their activity, their sense of agency, or “the power within”’(Kabeer, 1999:438). In Kabeer's view, since the different dimensions of empowerment are interrelated, a discussion of the meaning of one indicator inevitably makes reference to the others.

In making a strong case for viewing the different dimensions as being interrelated, Kabeer (1999) suggests it is not necessary to use the attainment of basic needs as an indicator of women's empowerment. However, this paper contends that in the African context this is indeed a milestone that has yet to be achieved among various social categories and therefore represents a level of empowerment. For example, African women's access to and participation in housing delivery in South Africa, after years of exclusion and marginalisation by the apartheid state, is an achievement that cannot be ignored and should therefore be used to determine the extent to which they have been empowered.

A second way of measuring empowerment in Kabeer's view is by using the United Nations Development Programme's Human Development Index, which is gender disaggregated, and the Gender Empowerment (GEM) Index, which goes beyond examining achievement in the basic needs to establishing achievements in areas such as women's representation in politics However, the GEM Index is criticised because it ignores women's choices and the values of their communities and uses a ‘definition of achievement which represents the values of those who are doing the measuring’ (Kabeer, 1999:440). Beteta postulates that the GEM Index fails to work in 60–90 per cent of low-human-development countries, and is therefore biased not only towards the developed countries but also towards the better-off women in the developing countries (2006:223). While noting that the legal and regulatory framework, women's social and economic rights, and societal attitudes are all key elements in women's empowerment, Beteta asserts that ‘no measure or composite index aggregating these dimensions has so far been suggested’, and recommends the construction of a ‘composite index measuring the existence or absence of a “GEEE” [Gender Empowerment Enabling Environment]’ (2006:236). The GEEE incorporates elements absent from the GEM Index, such as a legal and regulatory framework, and includes the protection of women's rights and the socio-cultural attitudes towards gender equality and the strength of the women's movement.

Commenting on a study that sought to examine women's empowerment in rural Bangladesh, Hashemi and Schuler observe that using quantitative methodology alone to examine empowerment presents various problems. They postulate that ‘Developing valid reliable measures of women's empowerment [is] one of the most difficult tasks … Behaviours and attitudes that might be used to measure women's empowerment in one society may have no relevance in another’ (Hashemi et al., 1996:637). To avoid ambiguity, they asked very specific questions. Indicators of empowerment were constructed from the responses and from direct observation, personal interviews and baseline data.

Hashemi et al. (1996: 637) argue that ‘the concept of women's empowerment is elusive and the potential for structured surveys to contribute to understanding of it is inherently limited’. The present paper concurs with Hashemi and Schuler that, at best, the measures can only partially capture the phenomenon of women's empowerment and the methodology used to collect the data in the case studies on which this paper is based underscores this position. Ndinda (2002a) discusses the multi-relational linkages approach (MLA) by examining a single case study, Glenwood II (Q-section), arguing that the extent of women's involvement in each of the levels constitutes their level of empowerment in that community.

4. CASE STUDIES IN URBAN KWAZULU-NATAL

The case studies described in this paper were based on the concept of self-help housing, which has been used in developing countries as a solution to the housing crisis in African cities such as Durban, Johannesburg, Nairobi and Lagos. Self-help takes various forms: unaided, state-aided, state-supported and conventional housing development, which all focus on household initiatives and the role of the state in supporting them in housing development (Harms, 1992). The shortcoming of these categorisations is that they overlook the involvement of external agencies such as non-governmental organisations (NGOs) in helping households to access housing, an approach that was widely used in Latin America in the 1980s and applied in South Africa in the 1990s by NGOs such as the Built Environment Support Group (BESG) and other NGOs affiliated to the Urban Sector Network (BESG, 1998).

The NGO approach to mutual self-help housing had various characteristics. First, participating households were provided with advice on housing design, site layout, technical and legal aspects of housing delivery and social development. Also included was support in terms of progress monitoring, to enable the community to manage the project and control the quality so as to ensure that technical problems were rectified and costs reduced. Training as part of the support provided by the NGO included construction, management, conflict resolution and negotiation skills. The NGO approach to self-help housing initiated by NGOs in South Africa resulted in what became known as the people's housing process, where the elements of support were promulgated in housing policy (NDoH, 1998). The case studies described in this paper were implemented with the support of the BESG through the delivery system known as mutual self-help, defined as ‘where a group of households provides housing for themselves and where there is an element of collective labour involved’ (BESG, 1998:8). The BESG argues that ‘The participation of households as a group is what distinguishes mutual help housing delivery from other forms of community based housing delivery’ (1998:8). Self-help housing in whatever form relies heavily on community participation.

The notion of empowerment discussed in the Urban Sector Network literature is limited to the project cycle, which results in what has been referred to as the ‘projectisation’ of development. The BESG and other organisations affiliated to the Urban Sector Network viewed empowerment deriving from the mutual help approach in terms of improved self-confidence, acquisition of shelter, and social and economic development (BESG, 1998). While these aspects represent notions of empowerment, they do not concretise what is meant by social or economic development in the specific context in which they were applied. The data in this paper shed light on the extent of women's empowerment through participation in housing development. The discussion of empowerment in the various activities leads to what the paper later describes as the multi-relational linkages approach to empowerment, as opposed to the project approach, which views empowerment through the lens of involvement in the project, whereas the MLA suggests there is need to critically examine the housing development process in order to understand the extent of empowerment more generally, and that of women specifically.

Data presented in this paper were collected as part of a larger study conducted in KwaZulu-Natal in 2000. The study covered five areas, two in Durban (Ezilweleni & Luganda) and three in Pietermaritzburg (Q-Section of Glenwood II, Nthutukoville and Thembalihle), among the African (Zulu) population where housing delivery was taking place. Ezilweleni and Luganda are in the Inner West Council of the Durban Metropolitan region (Figure 1). The patterns of residential mobility here show that residents had lived on their land as squatters and tenants of Indian landlords. The threat of eviction in the early 1990s triggered off the struggle for security of tenure. Ezilweleni and Luganda benefited from in situ upgrading through Independent Development Trust funding before the inception of the current capital subsidy scheme implemented in 1994.

Settlement in Nthutukoville, Q-Section (Glenwood II) and Thembalihle (Figure 2) was through land invasions and was motivated by the need for refuge from political violence in the early 1990s. In situ upgrading was carried out in Nthutukoville by the residents with the assistance of the BESG, which acted as a facilitator in the development process. Besides the land invasions, the Transitional Local Council, in their de-densification and slum clearance from areas around Pietermaritzburg, resettled squatters in the newly serviced sites, such as Thembalihle. The delivery system in Q-Section and Thembalihle was a mix of in situ upgrading and greenfields development, where services were provided in the area that was already settled, and greenfields development also occurred in the remaining area where core services such as water, tarred roads, electricity, water-borne sewerage and storm water drainage were provided.

In each area, two focus group discussions were conducted, one with women and another with men, each comprising six to eight participants. The household questionnaires were computed using the Statistical Package for Social Science, and descriptive statistics showing the demographic and socio-economic characteristics were generated. The data from the focus group discussions and individual interviews were recorded and analysed to discover the key themes: participation in housing delivery, access to building materials, training in construction and managing the delivery process in general. This paper focuses on one of the key themes that emerged from the study: empowerment.

Figure 1: Location of Luganda in the eThekwini municipality

5. EXTENT OF EMPOWERMENT IN HOUSING DELIVERY

5.1 Advancement into leadership positions

We had no idea of development during apartheid. Participating in the project has developed our minds. We are now confident in giving our views. Some of our members have gained construction skills. We have developed leadership skills because it is the members of the community running our Trust and Project Office and not outside developers. (Committee member, women's focus group discussion, Ezilweleni, Durban, 21 March 2000)

As leaders, women were involved in planning for services in the communities in various capacities; for example, treasurer (Q-section), secretary (Ezilweleni, Nthutukoville and Thembalihle), vice-chairperson (Thembalihle) and committee members (all areas). In contrast to the situation in Ezilweleni, Luganda, Nthutukoville and Thembalihle, the secretary in Q-section was a man and the treasurer a woman. Women leaders in the case studies represented women's interests to varying degrees. In Ezilweleni, women insisted on a transparent method of site allocation and this is what occurred. In Nthutukoville, women refused to move out of the invaded area and lobbied for the recognition of their right to secure tenure. In Q-section and Thembalihle, the level of community participation in decision-making was low, thus the extent to which the women represented the interests of fellow women was determined by the level of community participation in all the phases of project development.

Figure 2: Location of Nthutukoville, Thembalihle and Glenwood II around Pietermaritzburg

As Table 1 shows, most decision-makers in all of the areas sampled in this study were men; men constituted 55 per cent of the decision-makers, compared with 45 per cent women. Although the gender disparity in community political leadership appears to be evening out, the dominance of men in the leadership is conspicuous. The narrowing of the gender gap in leadership structures may be understood by looking at the political context in South Africa, which through the legislation of gender equality has led to the greater representation of women in leadership positions.

Table 1: Gender analysis of decision-making structures

Area	Men	Women	Total
Glenwood II	5	4	9
Ezilweleni	6	4	10
Luganda	6	6	12
Nthutukoville	5	3	8
Thembalihle	6	6	12
Total	28	23	51
Percentage of the total	55	45
Source: Author's survey, 2000.

The areas cited in this paper show that the current crop of women leaders have emerged from a period of liberation struggle where they fought and resisted removal by the local authorities (Nthutukoville) and eviction by landlords (Ezilweleni and Luganda) alongside the men. Advancing to leadership positions seems to have been a transition from activism to recognition of the roles they have always played. Women's representation in decision-making reflects the broader discourse of gender equality as enshrined in the democratic constitution. However, gender equality appears to be practised within a cultural paradigm of male dominance and women's subordination.

In all of the areas studied, no woman held the top post of chairing the development committee. The secretary portfolio was held by women in all the areas except Luganda. Only in two of the five areas (Q-section and Luganda) were the treasurers women. The uniform pattern of male dominance in top decision-making positions in all five communities can be partly explained by looking at Zulu culture, where men are considered the key decision-makers and women are not expected to express their views (Ngcongo, 1993). Yet this expectation constrains women's participation in key decisions that affect them as household heads and community managers. The failure to express their opinions resulted in the wrong choices and cost whole communities the top structures in Nthutukoville, Q-section and Thembalihle. Women attained leadership positions but the gendered cultural notions and practice kept them in lower, less powerful positions.

The present paper recognises that the dominance of men in leadership is not unique to Zulu culture but is a global phenomenon. King (1995) argues that although women are represented in organisations, the rate at which they enter executive positions compared with ordinary management levels remains low. Male dominance in top leadership and executive positions is encapsulated in the notion of the ‘glass ceiling’, a term that refers to ‘a barrier so subtle that it is transparent, yet it prevents women and minorities from moving up in the management hierarchy’ (King, 1995:69–70). This paper argues that, while culture has a role to play in structuring social relations and institutions, the underlying gender ideology in the specific culture explains the subordination of women and men's dominance in community and regional governance.

5.2 Acquisition of construction skills

Women comprised 36 per cent of all those who had received training in housing construction, compared with 74 per cent of men (Ndinda, 2002b). The trained women built houses from the foundation up to the roof – this was particularly evident in the mutual self-help housing project in Nthutukoville (see Table 2). Working alongside the men, women dug the foundations, cast the concrete, laid the bricks and put up the roof. Both men and women in Thembalihle were trained in bricklaying and plumbing but only the men were trained in skilled construction, the reason being that men, unlike women, had prior construction experience – along with the stereotypical urban notion that building was men's work. Yet women emphasised that they were ready to do any type of work as long as it provided them with income. ‘Those who did plumbing are getting work in the community because people are calling them to do work in their homes’, said a single female domestic worker in a focus group discussion in Thembalihle (Pietermaritzburg, 30 July 2000). By participating in the delivery process, the residents of Thembalihle – and women in particular – gained the confidence not only to participate in housing-related trades but also to venture into other areas where they saw opportunities to improve their status. Although few women were involved in the actual construction, they confessed that ‘we never thought it could happen in life’ (elderly female pensioner, mid-sixties, women's focus group discussion, Ezilweleni, Durban, 21 March 2000).

Perceiving that they did not stand a chance when competing with men for construction-related jobs, women argued that ‘people will always choose men’ (woman involved in bricklaying who had reverted to sewing after the construction of her house, women's focus group discussion, Ezilweleni, Durban, 21 March 2000). Women acknowledged that they were better workers:

But women are more reliable; they don't take smoking breaks, they don't drink, and they do work properly because they have children to feed with their money. They know that they are not popular so they do their work well. (Middle-aged woman, mother of four, whose house was built by a local builder, women's focus group discussion, Ezilweleni, Durban, 21 March 2000)

This perception points to the factors in favour of women's participation in housing-related trades, but this realisation alone is insufficient if women do not maximise such strengths.

Table 2: Gender analysis of training in housing delivery

Trained	Men	Women	Total
Yes	21	12	33
No	40	55	95
Total	61	67	128
Source: Author's survey, 2000.

5.3 Participation in the supply of building materials

In the areas sampled, 96 per cent of the residents had purchased their building materials from formal suppliers and a mere 4 per cent from the local women's groups. As the case studies indicate, certain hardware stores were selected to supply the communities with the building materials. The Thembalihle committee directed the community to purchase their materials from a hardware store outside the community rather than from women brickmakers in the community who were also subsidy beneficiaries. In Luganda and Ezilweleni, where the building materials business was dominated by formal suppliers, mainly Indian businessmen, the lack of participation in the sector by subsidy beneficiaries led the residents to purchase locally available materials, such as sand, pointing to the exploitation of end-users by building material dealers.

According to the women in Ezilweleni there has been a ‘great change’ in their lives: ‘We can do things on our own as individuals and as a community’ (middle-aged woman, early forties, owner of telephone bureau, women's focus group discussion, Ezilweleni, Durban, 21 March 2000). It is notable that in these settlements, through saving schemes (stokvels), women were involved in the production and distribution of bricks. In Nthutukoville, 17 women were involved in the production of bricks that they used to build their houses. In Q-section and Thembalihle, about 30 women were involved in brickmaking.

The condition that beneficiaries had to procure their materials from licensed suppliers, which resulted in a bias against informal suppliers, points to the power of donors over development outcomes. The participation of women in the brickmaking held the greatest potential for their economic empowerment and local economic development. However, since they were operating as informal traders, they could not benefit from the subsidy grant, which was only to be spent at licensed hardware stores. While the women material suppliers still got cash purchases, securing the contract to supply the community with bricks would have greatly boosted their business. The non-flexibility of the subsidy design advantaged the formal material suppliers and disadvantaged the informal (women) suppliers who were making bricks in Q-Section and Thembalihle. The voucher system that was adopted to guard against fraud is not free from abuse by interested parties, hence the need to rethink and restructure the issuing of the housing subsidy to ensure that beneficiaries do not end up being disadvantaged even when the state has put in place policies to benefit women, such as preferential procurement policies (RSA, 2000).

5.4 Qualitative impact of home ownership among women

The success of their mobilisation and collective protest empowered women individually and collectively because they achieved their goal of being recognised as legal occupants of land and housing in the five settlements discussed in this paper. Had these women been in a rural area, they would not have secured land tenure. In rural areas, land is held in trust for the community by the chiefs and indunas (village headmen). Despite the change in the Constitution that grants equal rights to both men and women, women's access to land in rural areas is constrained by the practices of the chiefs and indunas, who insist on allocating it according to Zulu custom, and that implies only men can access land (Meer, 1997). The accidental move to an urban area, administered by the local authority and not the chiefs, worked to the advantage of single mothers of Nthutukoville. Furthermore, it gave them confidence as individuals and as a community. This is illustrated by the phrase that ‘women have power now’ (middle-aged woman, Development Committee member, women's focus group, Nthutukoville, Pietermaritzburg, 6 May 2000).

As a woman in Luganda said:

We have better houses now … In the old government we had township houses for rent. The government gave subsidies for people working for the government only. The new government gave us subsidies for two rooms and we don't have to rent any more. I've got my own house now. (Young single mother, one of those invited to settle in the area after the subdivision, formerly living in Umlazi Township, women's focus group discussion, Luganda, Durban, 2 April 2000)

This shows the impact of subsidy housing on the lives of the end-users, especially women. Whereas in the past they could not access housing because they were unemployed, now they had access to housing regardless of their marital and occupational status. As the statement confirms, past housing policies were not based on welfare but rather on employment, and only those who had income were given houses. Despite their irregular incomes, women had to rent accommodation, which drained their meagre resources without giving them security – because it meant that in the event that they were unable to pay, they would automatically be evicted. In the same area (Luganda) there were some women who argued that the housing subsidy was ‘not sufficient to build a house. You get two rooms for children and parents. We need about six rooms for children and parents’ (middle-aged mother of four, women's focus group discussion, Luganda, Durban, 2 April 2000).

The small size of the houses was seen as a constraint to empowerment. As an elderly woman argued, ‘We can't rent the houses out because they are too small even for the families’ (member of sewing group and committee member, women's focus group discussion, Luganda, Durban, 2 April 2000). The starter units in Luganda initially comprised two rooms, but for those whose subsidies were approved after 1996 the materials were just sufficient for one room, owing to the inflation in the economy as a whole.

For the women in Nthutukoville, Q-section of Glenwood II and Thembalihle, ownership represents security from political violence, the space to think and be creative, and an improved quality of life. Empowerment through the control over resources means more than just having a roof over one's head. Some women were already sub-letting, having control over tenants, in contrast to the previous situation where they were subordinate as tenants themselves. This letting was mainly done by single women for extra income, as it was ‘better than being a prostitute’ (married woman, in mid-thirties, owner of a tuck shop, women's focus group discussion, Ezilweleni, Durban, 21 March 2000) – a profound statement that subsidised housing provision has liberated women from the indignity of engaging in exploitative and demeaning work.

The women's words and actions encapsulate the notion that empowerment is the realisation that they are in charge of their lives, as suggested by the Beijing Women's conference (1995) and illustrated in the women's own words:

Kakuulu! I feel well. Happy. I was having a big sore here [pointing to her heart], as the place where we came from, our houses were burnt down. I was not dreaming. But now, I'm dreaming about this house. I was not dreaming because of the violence. Now I dream as if my house is still there. Our former houses were destroyed in the political violence between the ANC and IFP. (Middle-aged woman, Development Committee member, women's focus group, in Nthutukoville, Nthutukoville, Pietermaritzburg, 6 May 2000)

The extract is illustrative of empowerment as ‘power from within’. As the respondent indicates, she was bitter because her home was destroyed. Her subsidised house has comforted her and given her refuge from violence, and she is now able to dream. The idea of being able to dream is consistent with Mercado's (1999) view of empowerment as the ‘power to’. Housing has provided the space for women to reflect on their life experiences in search of shelter. It has given women security and stability and has restored their dignity and self-worth. The statements of women in this paper illustrate the qualitative impact of subsidised housing on their lives and show how housing has empowered them. Home ownership has changed economic power from being the preserve of men to include women. It signifies the start of a process towards achieving gender equality in the control of material resources.

However, the process needs to be strengthened by the extension of housing finance for consolidation. Of the women interviewed in this study, 43 per cent (28 respondents) had made housing improvements to the initial starter structure. About 43 per cent (27 respondents) of the 62 male-headed households and 48 per cent (32 respondents) of the 66 female-headed households cited low income as the main constraint to making improvements to their starter housing. This finding was strongly supported by the occupation categories, which suggested that 33 per cent (22 respondents) of the women were concentrated in the informal sector, compared with only 15 per cent (nine respondents) of the men. While low income was the main constraint to accessing housing finance among women, lack of access to credit was mentioned as the key constraint by the men.

Women in Ezilweleni noted that they had more housing space than when they lived in informal housing that had no electricity or water. The way the women mentioned space in the context of housing emphasises its importance in their lives. Previously they spent time sourcing water and fuel; now they have these in their houses. The time is now spent baking fat cookies, biscuits and foodstuffs that they sell at bus terminals in Pinetown and at the nearby schools. Consistent with the existing literature (Kusow, 1993; Selat, 1993), this study argues that housing is not just about accommodation but is the site of production for informal sector goods and services. Luganda's location in southern Pinetown next to an industrial area provided a ready market for the goods produced in the community. The range and intensity of informal sector activities suggest that housing as a product had a greater multiplier effect for women than their direct involvement in the construction process and that women's economic empowerment derived from their location in the post-implementation activities in housing.

6. LOCATING WOMEN'S EMPOWERMENT IN HOUSING

So far this paper has focused on various attempts to measure empowerment. A number of approaches such as the GEM and the GEEE have been used to measure empowerment in general. Analysts such as Rowlands (1997) and Hashemi and Schuler (1996), while refraining from quantifying empowerment, use qualitative approaches to determine empowerment, as does Kabeer (1999). Beteta (2006), in recommending the use of GEEE, proceeds from the premise that most measures used by the Human Development Index exclude variables that are important in determining women's empowerment in Third World countries. The analysts discussed focus on measuring women's empowerment in general, but the present paper has focused on women's empowerment specifically in the housing sector.

The discussion about women's participation and the extent of empowerment points to the indicators of empowerment in housing development. The key levels of empowerment that emerge from the discussion are: leadership (decision-making), construction, involvement in the materials supplies sector, home-ownership, provision of social utilities, and post-implementation activities (Figure 3). Decision-making affects the choice and level of services and the final outcome in housing development. Women's participation at these different levels constitutes their empowerment. The indicators are illustrated in a framework that this study refers to as MLA. Whereas Ndinda (2002a) used data from one case study in discussing the MLA, this paper draws on data from a number of case studies and recommends the approach as a useful tool for understanding the level of women's empowerment in housing delivery. The multi-relational view of participation is conceptualised as consisting of project planning and implementation as well as the related linkages in housing development.

Most of the indicators used in discussing women's empowerment draw on a range of sectors, but the MLA focuses on a specific sector (housing) and examines the position of women in that specific sector. The level of participation in all the components of a specific sector of the economy is what constitutes their empowerment. Although the MLA might seem to differ from approaches have been used to determine women's empowerment, the approach is similar to that of Kabeer (1999) in that empowerment is established by examining relationships among the participants.

The MLA suggests that to understand women's empowerment in housing delivery it is necessary to deconstruct and establish the position and proportion of women in each component of the housing sector. Empowerment in each component of the sector can be seen as either high or low. There is a high level of empowerment when women are involved in a specific aspect of housing and when their contribution in the specific component is central. For example, women can be said to be empowered when they comprise a large proportion in construction and when their participation includes designing houses, establishing quantities, contracting and building. Conversely, there is a low level of empowerment when only a few women are involved and if their involvement is only in the unskilled aspects such as providing labour.

At a macro level, there is a high level of empowerment when women's participation includes all aspects of the housing sector. This implies that a substantial proportion of women are involved in policy-making, construction, material supplies and the downstream industries of housing such as furniture and fittings. The high participation of women in each of these components of the housing sector is what constitutes their empowerment. This paper therefore recommends the MLA as a framework for understanding women's empowerment in housing delivery. The model can also be used to establish the level of empowerment in other sectors of the economy, such as education, energy and finance.

Figure 3: Multi-relational linkages approach

Consistent with previous studies, this paper argues that empowerment is a process and not an outcome and it does not attempt to construct an index for measuring women's empowerment. The discussion and analysis of empowerment is confined to the housing sector in post-apartheid South Africa and the MLA as recommended in this paper attempts to highlight areas where women's direct participation would lead to their economic empowerment as defined in various policy documents that aim to bring about gender transformation in South Africa.

7. CONCLUSION: IMPLICATIONS FOR WOMEN'S EMPOWERMENT IN PRACTICE

The present paper has examined notions of empowerment and how different scholars have conceptualised the idea. While the authors generally agree that empowerment is a fuzzy term, there are differences in how the concept should be measured and this is shown by the various indices that have been developed to measure women's empowerment. These are general, whereas this paper argues that, to understand the level of women's empowerment, it is necessary to use a sector-specific approach.

While empowerment is explicitly stated among the intentions of post-apartheid South African housing policy, this intention is not clear in practice. It does, however, emerge as a by-product of participation in housing delivery, as a result of which women have gained leadership skills and been able to influence policy at the community level. Although men continue to dominate where construction is concerned, some women have gained construction skills and others have begun to supply building materials in their own community. The acquisition of skills in these areas represents the level of women's empowerment, but it is the acquisition of housing that has had the greatest impact on women's lives. Only by listening to their voices can one appreciate the role of housing in the empowerment of women in urban Kwazulu-Natal. Although Kabeer (1999) argues that the acquisition of basic needs is not necessarily an indicator of women's empowerment, the statements of women discussed in this paper suggest differently. Without housing, women remain at the mercy of those who provide for them; with it, they now have power, and they can dream, create, and plan their lives.

Notes