A successful slum upgrade in Durban : A case of formal change and informal continuity

In situ slum upgrades implemented through community participation are widely considered global best practice in efforts to significantly improve the lives of at least 100 million shack dwellers. This paper scrutinises the process and impact of community participation in a slum upgrade in Durban. Based on data from an ethnographic study of Zwelisha, a newly upgraded settlement north of Durban, South Africa, this paper presents a nuanced analysis of the upgrade process and the role of community participation in achieving successful outcomes in terms of significant improvement to tenure security and wellbeing (as defined by Zwelisha’s residents). The analysis shows that successful outcomes are intrinsically tied to the manner in which the upgrade process is implemented. The paper argues the formal changes that result in successful outcomes are possible only because of informal continuities; specifically, the continued and consolidated power and influence of the local community development committee following upgrade. While the paper frames the continuity of informal power relations as important to successful outcomes for wellbeing, the findings can also be read as a tale of caution of how the state’s approach to community participation in slum upgrades may consolidate and legitimise informal power relations that may not be necessarily benevolent. 2013 The Author. Published by Elsevier Ltd. Open access under CC BY license.


Introduction
The active participation of slum residents in upgrade programmes tends to be captured by the term 'community participation' and has been lauded as global best practice in slum upgrades. 1 This paper deconstructs 'community participation' and its role in the in situ upgrade of a settlement in South Africa. It explores the impact of community participation on the relationship between those who directly participate in the upgrade process on behalf of all residents (i.e. the community development committee e CDC), and other (non-committee) residents. The paper argues that the relationship between committee and non-committee residents throughout the upgrade process is imperative not just to the successful implementation of the process itself, but what happens after e the maintenance and upholding of rules on planning and construction, approaches to home maintenance and improvements to self-conceptions of tenure security and wellbeing. It is in this context that the paper claims the upgrade was successful and that a particular type of community participation was essential to that success.
This paper presents an in-depth case study of the settlement of Zwelisha, 35 km north of Durban within the municipality of eThekwini. The findings cannot be generalised to beyond Zwelisha, but the insights offered by this paper on the intricacies of community participation carry a wider relevance to practitioners and academics engaged in work on slum upgrades.
The paper is structured in five parts: the first part presents an overview of literature on community participation and slum upgrading, and applies this to the South Africa experience, specifically in eThekwini Municipality. The second part presents the research methodology. The third part introduces Zwelisha and the relationship between its leadership and residents. The fourth part presents a nuanced analysis of how community participation plays out in the upgrade of Zwelisha. This part also discusses the concept of 'success' and its application to the case study. The fifth part concludes the paper by drawing links between successful outcomes from formal change and the continuity of the informal CDC, thereby deepening current literature on community participation and slum upgrading.

Context of community participation in slum upgrades
Much has been written on the theory and practice of community participation in housing programmes in the Global South. A great deal of this literature from the mid-1990s onwards assesses the practice of community participation against Choguill's 'Ladder of Community Participation' (1996), itself based on Arnstein's (1969) 'Ladder of Citizen Participation'. Choguill's ladder classifies eight types (or rungs) of community participation from empowerment (the highest level and gold standard) to self-management (the lowest level that represents a neglectful relationship between state and citizen). To Choguill, community participation has two roles that can be assessed against this ladder: it is a means to achieve basic needs such as housing or infrastructure; and it is a means for communities that tend to be marginalised to influence decision making, and thus politics in housing development. This distinction between meeting basic needs and influencing politics is useful to assess community participation in the upgrade of Zwelisha.
There have been many adaptations of Choguill's ladder in assessments of housing provision to factor the multi-faceted character of 'community participation' and its increasingly broad application in housing interventions (e.g. Davidson, Johnson, Lizarralde, Dikmen, & Sliwinski, 2007;Lizarralde & Massyn, 2008). Community participation can relate to processes and the engagement of certain sections of society in the design, implementation and evaluation of housing interventions. It can therefore be a means to achieve basic needs or political influence (as Choguill states), or it can be an end in itself. For example, in the South African context, community participation in the design and delivery of local development is enshrined in the post-apartheid Constitution and is regarded as a means to redress the historic imbalance of power (Williams, 2006). Thus an upgrade process with a heavy community participation component may aim to and actually uphold Constitutional will, without necessarily meeting housing needs and/or improving the political standing of citizens in housing or other spheres. Often though, community participation as means and ends is more closely related. Participation can be a part of broader social goals, for example, in politically repressive regimes organising marginalised citizens and having them demand and manage resources as part of self-help housing can carry intended spill-over effects into other spheres (Stein, 2008); or partnerships between slum dwellers and state actors in upgrading programmes, it is hoped, can gradually reconfigure conceptions of urban citizenship and the rights and responsibilities of the poor and the state (McLeod, 2011). Whether community participation is advanced as a means or an end, its potential to empower marginalised groups, as Choguill (1996) notes, is its highest achievement.
However, the case of Zwelisha demands we ask 'the empowerment of whom?' and 'to what effect? ' Lizarralde and Massyn (2008) found that the consequences of community participation and the engagement of some marginalised groups in decision making and planning negatively impacted other settlements nearby that were not invited to participate and thus helped to increase social fragmentation at a city-wide level. Mafukidze and Hoosen (2009) similarly found negative outcomes of community participation in their case study also in South Africa. They found the community participation process actually engendered greater distrust between citizens and the state as it raised expectations for housing that the state ultimately failed to deliver. Such findings have prompted a reassessment of the value of 'community participation' in low income housing as an unquestioned good with the potential to empower. This paper engages with these critiques but arrives at a different conclusion by assessing community participation against these outcomes: basic housing needs are met, political power has increased and residents' self-defined tenure security and wellbeing has improved.
To conduct this assessment, the paper uses Choguill's (1996) ladder of community participation in two related ways. The first, it is applied to assess the extent to which the process to upgrade Zwelisha is participatory, and the type of community participation it mostly resembles (below). The second, it is applied to understand and assess social relationships and power dynamics within the settlement (part 4, Results and discussion). With regards to the upgrade of Zwelisha it is useful to distinguish between three tiers of actors: non-residents with overall responsibility for the design and delivery of the upgrade (e.g. the municipality, housing consultants, builders, utility providers); the community development committee, a seemingly representative body with approximately 10 members (exact numbers were difficult to obtain) all of whom reside in Zwelisha 2 ; and ordinary residents in the settlement i.e. non-committee members. All three tiers participated in different ways in the upgrade process and adopted various models of partnership with each other throughout. It is the combination of these tiers of actors working in different elements of the upgrade process that result in housing, increased political power and improvements to tenure security and wellbeing.

Slum upgrading in eThekwini Municipality and the role of community participation
Since the publication of the state's housing policy 'Breaking New Ground' (GoRSA, 2004), in situ upgrading is the preferred approach to slum improvement. Under the National Housing Code (GoRSA, 2009: 15), the role of the municipality in slum upgrading is to act as "developer", that is "The municipality undertakes all planning and project activities". At the feasibility stage of the upgrade process, the Metro Housing Unit in eThekwini Municipality commissions various impact assessment studies, land acquisition studies, architectural plans and engineering surveys and so forth to determine whether an upgrade is feasible. It is a process that assumes the residents of slums want both upgraded facilities and the municipality's plans to deliver these upgraded facilities. If a site is recommended for upgrade Metro Housing appoints a project manager who is responsible for monitoring the quality of commissioned work during both the feasibility and implementation phases. Many aspects of the upgrade project are subcontracted to consultants, including sometimes the position of project manager (Patel, 2012).
The project manager is responsible for appointing a community liaison officer, and together they liaise directly with the local councillor and any settlement-based authority (in Zwelisha this was the CDC). The councillor and CDC are responsible for 'raising the voice' of shack dwellers to ensure some resident participation in the upgrade process. There are three major assumptions here: first, the CDCs, which are in principle elected, are founded and organised in accordance with just and democratic principles; second, that councillors act in favour of shack dwellers in their constituency, sometimes over the interests of their middle class formally housed voter base; and third, that slums are communities with a communal identity that incentivises residents to behave cohesively for the greater good.
2 At the core of the committee was a group of 3e4 men who are unchanged since mid-1990, across the committee, at the time of fieldwork, was a broadly even mix of men and women.
Under initiatives for greater community participation, settlement-level actors are very involved in the process of upgrading. Although they tend not to be (as in Zwelisha) involved in the design of houses, plans for the settlement's layout, approaches to the upgrade process or even in the evaluation of an upgrade. Instead, these actors tend to act as agents of implementation with specific responsibilities such as drawing up housing lists of eligible residents and monitoring any newcomers that arrive so they do not take advantage of the upgrade. They also facilitate the entry and movement of builders, engineers, water and sanitation officials and so forth during the feasibility and implementation phases of settlement upgrade. The role of settlementlevel actors is vital to the logistical success of an upgrade project (Taylor, 1994). However, there appears to be very little official and mandatory monitoring of how these actors operate. This suggests an under-acknowledgement of their power and influence over which individuals benefit from an upgrade and which do not.
Against Choguill's ladder, the participation of select members of Zwelisha in the feasibility phase of the upgrade closely resembles 'informing'. At this level, 'informing' denotes, "a one-way flow of information from officials to the community, of their rights, responsibilities and options, without allowance for feedback or negotiation, in projects that have already been developed" (Choguill, 1996: 439). In Zwelisha, the CDC is informed of their selection for upgrade and the manner in which the upgrade will be implemented. Choguill argues the 'informing level' is tokenistic as a form of community participation and is a form of manipulation. The municipality, by informing the CDC of what will happen exercises control over what change in the settlement looks like and its pace. During the implementation phase, however, the role of the CDC develops. For example, it is given specific roles to perform that include a highly visible role facilitating the entry and exit of professionals, and decision-making responsibilities over who is eligible (via the housing list). The relationship between the CDC and municipality during this phase resembles more of a partnership where some formal decision making is shared. Although it is a limited partnership in that the municipality pre-determines the arenas in which the CDC's decisions count.
The relationship between the CDC and municipality, when seen through a lens of community participation, is largely tokenistic with some elements of a partnership model as the upgrade process develops. It is an approach that yields the delivery of the upgrade project on time and on budget (according to the plans of the municipal project manager). Although, further explored in part 4 (Results and discussion), this is an incomplete assessment of the role of community participation in Zwelisha's upgrade. From the somewhat limited 'empowerment' garnered from the community participation approach, the CDC uses this state-sanctioned influence to powerful effect over ordinary residents, specifically, expanding its role in the upgraded settlement.

Methodology
The data this paper draws upon was collected and analysed as part of a wider research project into the effects of the upgrade and formalisation process on the tenure and vulnerability of shack dwellers in three different settlements in and around Durban (Patel, 2012). The data were collected over nine months between 2009 and 2010. An initial three month pilot study identified three settlements at different stages of the upgrade and formalisation process so that the process could be 'followed' (one settlement was at feasibility stage, one at implementation stage and Zwelisha was at post-implementation stage). At this stage of the study eighteen semi-structured interviews were conducted with professionals involved in the upgrade process including municipal officers, master builders, housing consultants and local councillors. Over a further six months, ethnographic data was collected to document changes in the lives of eight residents in these settlements. This involved five different rounds of interviews including in-depth semi-structured interviews, oral histories, social relations mapping, community mapping and the use of diaries to record acts of reciprocity involving the respondent.
In Zwelisha, five of the respondents were women, three were men, the youngest respondent was 23 and the eldest was 68, six identified themselves as amaZulu, one as Shona (from Zimbabwe) and one was South African of Indian origin. The earliest settler amongst the respondents had arrived in 1993, and the latest arrival moved to Zwelisha in the midst of the upgrade process in 2007. All the respondents lived in so called 'RDP houses' (houses built using a state subsidy) in the part of Zwelisha also known as Trenance Park 4B (there is also a 4A, 4C and 4D); seven were owneroccupiers and one was a tenant. They were identified according to the contribution of their characteristics to a mixed group of respondents in order to reflect diversity in the settlement (Barbour, 2008: 53).
Respondents were initially identified through close collaboration with a research assistant who lived in a part of Zwelisha that was not Trenance Park 4B. The assistant was an older Xhosa woman who had lived in the settlement for over fifteen years. All interviews were conducted in isiZulu or English depending on the fluency and comfort of the respondents. Interviews in isiZulu were translated into English at the time by the research assistant to enable the lead researcher to ask follow up questions. There was a danger that responses were influenced by the presence of a person known to the respondent, and that in only speaking to known contacts of the research assistant a heavy bias was introduced. However, as Rashid (2007) explores in her study of an informal settlement in Dhaka, the hazards of employing gatekeepers must be balanced by the need to gain access to research participants and the necessity to build trust. In this study, the trade-off between the possible influences of a research assistant was outweighed by the gains made in terms of access and trust. To help manage the inherent bias in the selection of respondents, following initial recommendations by the research assistant, further recommendations of people to approach were made by the respondent, with the same effect on trust-building.
Zwelisha was selected as a study site in consultation with a Senior Planning Officer at eThekwini Municipality against criteria including location, stage in the upgrade process and size (for consistency between all study sites in the larger research project). According to the municipal Project Manager responsible for the upgrade, Zwelisha was selected by the municipality because it promised to be a relatively uncomplicated project e the site was close to existing infrastructure and the land acquisition process had been straightforward. Additionally, there was sufficient land so that no resident would have to be resettled elsewhere, the settlement was small (513 households were living in RDP houses at the time of the fieldwork), party political activity within the settlement was minimal, the councillor was supportive of the process and widely respected, and despite the re-generation of much of Durban prior to the 2010 Soccer World Cup the municipality managed to secure a respected master builder to oversee the implementation phase of the upgrade (2005e2009).

Background to Zwelisha and its residents
Zwelisha is isiZulu for 'new land'. Its historic origin is contested by residents; their narratives link its history to contemporary competition for power and leadership. Competition has heightened since the upgrade of the settlement. The history presented here is based on common recollections of the study respondents, all of whom are former shack dwellers who now reside in RDP houses and most of whom describe themselves as ethnic Zulu.
The earliest settlers arrived in the mid-1980s from the old Transkei (now Eastern Cape) and were Xhosa-speaking. They cleared land at the foot of surround hills and established a settlement on the bank of a river. These settlers established the first leadership structure and oversaw who came to live in the settlement and under what terms. Many had arrived without identity documents or the correct apartheid-era permits to live in KwaZulu-Natal (KZN). By the end of the 1980s/early 1990s there was a growing presence of amaZulu in the settlement. These new settlers resented the influence and power of the Xhosa-speakers in a Zuluspeaking province. Violent confrontations were commonplace at this time in the settlement and across the province. Reflecting the changes to Zwelisha's ethnic demographic, since the mid-1990s the ethnic composition of the leadership altered. The new leaders brought with them new ideas about tenure security and housing that directly clashed with the status quo and preferences of the old leaders. The contemporary contest for leadership is framed by respondents as a struggle between 'new' and 'old' settlers. The 'new settlers' dominate positions in the CDC. Current leaders claim they were elected at a community meeting and that frequent community meetings are held where ordinary residents can voice their concerns to and about the leadership. The 'old settlers' include members of the old leadership from the 1980s and the descendants of early Xhosa-speaking settlers who tend to believe that their turn at the leadership was usurped.
By 1997, there was open tension and disagreement in Zwelisha over the issue of housing. The residents largely fell into two groups: one campaigned to the municipality for an upgrade and better housing, and the second was committed to maintaining shacks. This second group was commonly believed to have some vested interest in the status quo. The assumption amongst many Zulu residents is that Xhosa-speakers staked claims to the largest plots of land when the settlement was first founded and were dominant shack-lords with much to lose by settlement upgrade.
In 1998, the CDC lobbied the local councillor and obtained a commitment from him to support the upgrade of the settlement. The Zwelisha upgrade project began in 2005. The leadership of the settlement has not changed since the mid-1990s. The current CDC oversaw the entire upgrade process. Following the upgrade and widely perceived resource advantages that accompany it, the displaced or aspiring leaders (old settlers) have heightened their challenge for leadership positions and are attempting to recruit a support base amongst those still living in shacks in Zwelisha (those ineligible for an RDP house). Their numbers, however, are minimal compared to those residents living in RDP houses. The early ethnic divisions apparent in the leadership struggle in Zwelisha are not so obvious in the current power struggles. The people still living in shacks, for example, are not exclusively Xhosa or Zulu speakers. However, the early Xhosa-speaking settlers were largely without proper documents. Therefore, during the upgrade process their paperwork was more complicated than others, which typically delayed their occupation of an RDP house.
Since 2009 and the completion of all building work, other than a few remaining shacks at the edges of the settlement, Zwelisha appears as neat rows of pastel shaded houses with blue mail boxes lined up alongside tarred roads. Of particular note is that every RDP house looks exactly the same e there are no backyards shacks or outbuildings and no extensions that jar with the original building, any extensions are seamless; built in the same style, with the same materials and in the same colour as the original house. For an upgraded settlement in South Africa, the absence of shacks abutting RDP houses is surprising (see Lemanski, 2009).

Results and discussion
This part of the paper offers nuanced analysis of how the upgrade process played out in Zwelisha focussing on the role of the CDC and its relationship with ordinary residents. The CDC played three central roles during the feasibility and implementation phases of the upgrade: it developed a housing list, facilitated the entry, movement and exit of external professionals, and attended meetings with the municipality and the local councillor in its capacity as representative of Zwelisha's residents.
In order to develop a housing list the CDC organised all residents to attend a series of community meetings in the settlement. At these meetings the CDC informed residents, as they had been informed by the municipality, that the settlement of Zwelisha was to be upgraded and formalised. The CDC alone developed and delivered messages to other residents, and determined the timing of messages. To reach those residents who did not attend community meetings, members of the CDC canvassed the settlement to determine the eligibility of all residents to the process.
The process to develop a housing list of eligible residents was decidedly less contentious than in other settlements in the wider study. Criteria for eligibility to the housing subsidy are set by the municipality. 3 In the case of in situ upgrades, these criteria are often supplemented to include date of entry to the settlement (for example). Because many upgrades entail de-densification, the upgraded site usually cannot accommodate all residents that were living in shacks. In such cases, eligibility to a subsidy is not affected, but the ability of residents to capitalise on that subsidy and move into a subsidy-built (or 'RDP') house, is. In Zwelisha, no such supplementary criteria were employed, largely because there was sufficient space on which to build housing units (in fact Zwelisha was able to absorb eligible residents from other settlements who had been living in transit camps awaiting a subsidy-built house). The CDC's task was therefore to ensure all eligible residents applied for a subsidy. They did this by manoeuvring residents through every stage of the upgrade process, issuing instructions and actively helping residents to comply with these instructions. This micromanaged approach left little room for residents to understand and engage with the upgrade process at large, or make their own decisions regarding their participation in the upgrade.
One respondent, an elderly man who had lived in Zwelisha since 1996, recalled the moment he learnt about the upgrade process: "I had no involvement in the plans. We first heard about it at a community meeting. that's [also] when we learnt the house will be two rooms. At that meeting only, before that we knew nothing" (Interview A, 13/05/12). The same respondent adds, "The community committee [then] called people to the office, they listed [registered] my name and ID number for a house" (Interview A, 13/ 05/12). Another respondent, a woman in her mid-40s, added, "The committee [then] said they'll take our forms to the Housing Department" (Interview B, 14/05/10). Respondents did not have any direct communication with the Housing Department. After several months and once construction had started, "they [the CDC] called us to the office to give our number for the house. They told us to keep this number until a house is built. We [all] got different numbers." (Interview C, 13/05/10), said a male respondent in his late 30s. Once the first tranche of houses were built the CDC 3 At the time of fieldwork, the criteria for eligibility for a state subsided house in eThekwini were: (i) must be a South African citizen; (ii) must not have received government subsidy before; (iii) combined household income must not exceed R3500; (iv) must not own or have owned property before; (v) must be married or cohabiting with long term partner or single with financial dependents; and (vi) must be 18 years of age or above. Other types of subsidy were available for higher income earners. directed residents to their new houses. The same respondent recalled, "The committee was reading from a list when they gave us our keys and house number. Someone within the committee wrote the list. The houses are not all the same e some are built better than others. Some rooms have rain leaks inside, some have cracks. It's luck who gets what" (Interview C, 13/05/10). Another respondent observed that in addition to a list, "They [the CDC] had a map and showed you, 'you must move here' e there was no choice. The numbers were already written. We saw our [house] number on a map. Our name was next to the house number on a separate list" (Interview A, 13/05/10). Most respondents believed it was luck or fate that led them to occupy the house they do; there is no suggestion of foul practices at play.
Residents that did not attend community meetings were given special attention by CDC members. One respondent, a man of Indian origin, did not speak or understand isiZulu and so did not attend meetings. He recalled, "The people of the community came [to my home] and told me I must come and register my name and [ID] number to get a house, because they are changing this place" (Interview D, 13/05/10). He duly did as he was told. There was a potential complication however, in that his shack belonged to his step-father (who had left the settlement with the respondent's mother years earlier). The respondent did not have a good relationship with his step-father and there is little to suggest that he wanted his shack to pass to his step-son. Nevertheless, "They (the committee) told me to register for a house.
[Then] they changed his name for my name.
The second principal role of the CDC was to facilitate the entry, movement and exit of housing and services professionals in Zwelisha. The leader of the CDC and/or other CDC members would accompany engineers, construction workers and utility service providers wherever they went in the settlement in order to ensure their safety in an unfamiliar place, to be able to answer questions from residents about who these people were and what they were doing, and to learn from these professionals what exactly they were doing and why so that CDC members would be able to communicate this to residents at meetings. Observing these professional with CDC members, one respondent stated, "The committee helped us a lot to get these houses. I'm proud of my house. The municipality was not here for long e helping with roads, bins and water. They were here only for a short time. The first [time they] came [was] after we got our houses. I think because our place was shacks the municipality thought they are not allowed to help us. Only the contractor came to build the houses" (Interview E, 14/05/12). The respondent, an elderly woman in her late 60s, later explained why she believed the municipality was not involved in the upgrade, "The committee, they gave us the number to the houses, so I think it is them who helped us." (Interview E, 14/05/12) Most other respondents, however, recognised a greater role played by municipality during the upgrade and the role they continue to play in, for example, ensuring refuse is collected.
External professionals also recognised the major role played by the CDC. The master builder appointed by eThekwini Municipality to oversee construction recalled, "We had a small Greenfield site just next to the settlement. So we built fifty houses, told the community committee, then fifty families moved in. We demolished their shacks and built fifty more on that site. And so forth until all area was upgraded" (Interview F, 11/11/09). The role played by the CDC was essential to allow external professionals to progress with their work plans and fulfil their contracts with the municipality.
The third role the CDC played was to attend meetings with the municipality and local councillor in order to ensure the representation of Zwelisha's residents at that level, and to relay information back to residents. Through this position of gatekeeper, the CDC was able to manage channels and content of communication between Zwelisha's residents and the municipality. "The committee reported to us in meetings what was happening" (Interview E, 14/05/ 10), was a popular statement amongst all respondents. Avenues for either the municipality or residents to communicate outside of the CDC were almost non-existent, only if there was an issue with a resident's application for a subsidy did they communicate with a housing official. Such incidences were isolated.
To the local councillor and municipal Project Manager, a well organised CDC who delivered a complete and accurate housing list, was available to escort external professionals, and provided continuity in meetings, meant the upgrade process ran smoothly. These actors had little reason or incentive to investigate how the CDC interacted with ordinary residents and what 'community participation' looked like at the settlement level.
There were three main effects in terms of the relationship between ordinary residents and the CDC stemming from the CDC's role before and during the upgrade. The first effect was a widely held belief amongst residents that the CDC was instrumental in bringing the upgrade and its benefits to Zwelisha. This belief stemmed from the high visibility of CDC members with nonresident professionals, the ability of the CDC to speak with authoritative knowledge on the process at community meetings, and popular narratives on the history of the settlement and accession to power of the current CDC. The second effect was gratitude to the CDC, particularly amongst marginal groups within the settlement. For non-isiZulu speakers and anyone did not attend community meetings, the CDC targeted them and drew them into a process that they did not know about, and as a result they have a house. The respondent who was of Indian origin was clearly delighted with the outcome of the process; he said "I will stay here until the day I die. It's the first house in my life I've ever owned!" (Interview D, 13/05/10). The third effect, for those residents who knew of the involvement of state actors in the upgrade process, was awareness that the CDC's power had been recognised and (through their role in the upgrade process) legitimated by the state. This third effect plays out in CDC-resident relations postimplementation.
Following the completion of all building works in 2009, the CDC's role in settlement housing evolved. With their power and authority afforded greater legitimacy by the participatory nature of the upgrade process, the CDC effectively began to regulate land and house use post-upgrade. The CDC, claiming broad support from the municipality and local councillor, prohibited the building of shacks and poorly constructed extensions on land that is privately owned (or will be once title deed are issued) by Zwelisha's upgraded residents. A respondent explained, "Even though we have big yards you can't build a shack e no one is allowed. Only if you have space and build a proper extension are you allowed. That's what the committee and the councillor says" (Interview A, 13/05/10). The local councillor was involved in the upgrade process and worked very closely with the CDC and municipality throughout. He does not feature strongly in the narratives of respondents, although his support of the CDC appears to lend it greater legitimacy to enforce rules on land and house use. This respondent added that if the CDC learns of an extension or the building of a shack, they inform the municipality, "There are housing officers from the municipality who say, 'you have a two-room house and you can only extend not build other shacks'. They come and pull down the mjondolo [shack] if you've built one e even during the day when you're at work, or at night. It's happened to many people. They are always looking." (Interview A, 13/05/10).
These rules have greatly restricted livelihood opportunities for homeowners, particularly their chances to become landlords and engage in backyard rental. The CDC also banned shebeens [taverns] in the new settlement, arguing they attract unsavoury characters, encourage excess drinking, can become hubs of violent and aggressive behaviour, and that residents in the surrounding area suffer the consequences. These planning restrictions are intended to create a pleasant unthreatening and aesthetically pleasing environment. The CDC appears to be able to implement such rules because to an extent residents believe that these restrictions have positive outcomes (especially regarding shebeens), and because they risk losing their investment (in building materials) if they fail to comply. Through monitoring adherence to planning rules, the CDC exerts control over elements of residents' lives.
Post-implementation, the relationship between the CDC and municipality has also evolved. While all building work is complete and municipal meetings regarding the upgrade (to which the CDC is invited) are less frequent, the CDC is able to maintain a relationship with the municipality on a different basis. Not only does the CDC call municipal officers if a shack has been built, but CDC members are engaged with municipal officers to help residents who were unable to access the housing subsidy the first time around (usually because of missing paperwork). One respondent observed, "There are some people in the shacks still so the community committee is talking to the municipality to try to find them a place to stay in a house" (Interview A, 13/05/10). The CDC has identified roles, beyond those initially identified for it by the municipality and local councillor, that increase its ability to access formal political power and influence.
Reflecting on the upgrade process, from the perspective of a broad range of (committee and non-committee) residents living in Zwelisha, the upgrade was a success on the basis that basic housing needs were met, political power increased (for the CDC) and residents' self-defined tenure security and wellbeing improved. Tenure security, broadly defined as a resident's ability and willingness to reside in a given space (Patel, 2013), is an integral part of an individual's wellbeing. It is based on factors that include how safe a resident feels in and around their home, how secure they are in their claims to occupy land/shelter and their ability to access appropriate and trusted levels of authority to arbitrate occupancy claims in the event of a dispute. Significant improvements to tenure security and wellbeing (as noted by residents) are directly attributable to the upgrade and the CDC's role in its implementation. For residents, having a house and access to utilities on land that has been formally identified as theirs, and roads, street lighting and refuse collection throughout much of the settlement, means a better standard of living. Also, there has been a significant reduction in intrasettlement violence at two levels: the first, the settlement leadership level. Because of the consolidation of the CDC's power and authority following its role in the upgrade, life in the settlement is calmer. After years of uncertainty and violence in power struggles, the position of the current CDC is virtually unassailable. The efforts by 'old settlers' are broadly regarded as weak and their power is waning. Their support base is only the few who still live in shacks and the CDC is helping them to obtain a subsidy-built house. After more than twelve years, there is stability in the settlement. The second level is between individual residents. The absence of shebeens and the presence of a strong leadership, who all respondents felt they could approach to arbitrate disputes, appear to have reduced conflict, particularly violent conflict, in the settlement thus enabling residents to feel safe in and around their homes.

Conclusion
This paper argued that Zwelisha is an example of a successful slum upgrade because of the manner in which the upgrade process was enacted, which meant that formal changes in the settlement that led to improved security of tenure and improved quality of life (defined by residents lived realities) depended upon informal continuities; specifically the consolidated power of the local community development committee whose core membership has not changed since the mid-1990s. Continuity in this context has another dimension: the relationship between formal state actors and former 'informal' citizens is still mediated by the CDC.
Turning to Choguill's ladder of participation, where relations between the municipality and the CDC, assessed through a lens of community participation, can be regarded as manipulative with only some elements of a partnership model that emerge as the upgrade process develops. Relations between the CDC and ordinary residents, when also assessed through a community participation lens resemble a compound of manipulation that includes elements of 'informing' and 'conspiracy'; the former where a one-way flow of information dictates the rights, responsibilities and obligations of residents, and the latter where, "no participation in the formal decision-making process is allowed or even considered" (Choguill, 1996: 439). In Choguill's description of 'conspiracy' she explains communities are not permitted to participate because the government appears to reject the idea of helping the poor. This caveat does not appear true of the CDC whose prolonged effort to ensure residents obtain a housing subsidy (for example) suggests a benevolent motive in their manipulation of Zwelisha's residents. Rather than seeing this manipulation as a negative consequence of community participation, assessed against criteria of secured access to housing, improved tenure security and wellbeing, it is a success. Furthermore, the continued engagement of the CDC in land and house use suggests a sustainability of upgrade outcomes. Political empowerment is more difficult to assess. The political empowerment in this process concerns the CDC, not ordinary residents. The CDC has expanded its influence and reach both up (to the municipality over time) and down (over residents). But at what cost in the future? Given the role the upgrade process has played in consolidating authority and conferring legitimacy on the current CDC, if the majority of residents wished to change the leadership would this be possible? At the moment it is not possible to answer this important question, but it is worth factoring in to future assessments of Zwelisha's upgrade process and its long term success.