New housing/new crime? Changes in safety, governance and everyday incivilities for residents relocated from informal to formal housing at Hammond’s Farm, eThekwini

New state-subsidised ‘RDP’ housing in South Africa aims to provide former informally-housed residents with a better quality of life, stronger community and decreased levels of crime. Despite the state’s ambitions, this process is highly contradictory, increases in safety occurring alongside rising incivilities and tensions. This paper contributes to an emerging set of debates on the socio-political outcomes of state-led housing interventions in the global South, through an illustration of the limitations of efforts to produce ‘safe neighbourhoods’ in contexts of high unemployment alongside high levels of violence. The conceptual framing of ‘Southern Criminology’ (Carrington et al., 2015), centres the significance of histories of colonial and post-colonial violence, inequality, hybrid governance and justice practices, as well as informal living, and is employed to analyse recently housed residents’ experiences of crime and safety in South Africa, in a north eThekwini settlement, Hammond’s Farm. Recognising these ‘Southern’ factors, the paper argues that movement into new formal housing, is typified by significant material changes at the home and neighbourhood scale which foster privacy and safety, formalised governance practices and (partial) improvements in policing services. These occur in conjunction with access to new leisure activities including alcohol consumption and ‘township life’ which alongside ongoing poverty foster urban incivilities. A ‘Southern Criminology’ perspective frames concluding questions about the nature of crime in contexts of urban change, which are persistently shaped by inequality and wider historical and structural factors, challenging the state’s aspirations to achieve crime reduction through housing.


Introduction
This paper's contribution is to examine the relationship between rehousing poor residents in a South African city, and their consequent experiences of crime, broadly defined. Its origin lies in a British Academy funded project (2014)(2015)(2016) examining the intersections of dynamic socio-spatial relations in urban South Africa (and urban India), namely: changing housing (as a political, economic and material construct) through state provision of formal privately-owned properties and the subsequent related social practices, domestic violence in particular, which emerge in the context of such socio-spatial change but are not necessarily 'caused' by such changes. Using a mixed qualitative methods approach, this broader premise acknowledges the complex connections between changing spatial relations and social processes, and places the specificities of context and wider structural realities as central to interpretations. This particular paper focuses down on three specific contradictory crime-related findings for resi-dents who have been relocated into formal housing, in the settlement of Hammond's Farm, in the city of eThekwini (Durban) in South Africa. Gendered outcomes, including domestic violence, were also evident, but are not the focus of this paper (instead see Meth and Rajasekhar, 2016). The three findings evidence a broad approach to 'crime' and reveal the following: that relocation into formal housing does impact on residents' privacy, safety and space in significant and often positive ways, that wider changes to governance structures and practices are discernible, and finally, that 'new' incivilities associated with, what is here called 'townshipization', are evident. Historical and context-specific factors are key, and the paper concludes that histories of living informally, and ongoing poverty and unemployment remain central to experiences of crime in a context of housing change.
In order to account for South Africa's nexus of high crime levels, entrenched vulnerability, the informal-formal housing context, fluid systems of governance and the political significance of a housing welfare intervention in a fraught context with a history of racial discrimination, the paper draws on the conceptual framing of 'Southern Criminology' (Carrington et al., 2015), and related framings (Southern Planning Theory, after Watson, 2014) to analyse the findings. The turns to Southern theorisations follow Connell's 2007 critiques of how global social science knowledge is shaped by various divisions so that European and North American voices, theories and experiences dominate, to the exclusion of those from outside the 'metropole' (Connell, 2007). Carrington et al. (2015) use Connell's theorisation to propose a 'Southern Criminology' aimed at adding "new and diverse perspectives to criminological research agendas" (2015: 2). They identify characteristics of the global South which justify such a theoretical position: namely; where nation states do not commonly exhibit "high level[s] of internal peace", where state formation is shaped by colonial histories (introducing racialized practices of marginalisation and injustice), where violent phenomena are significant, where forms of justice discriminate against indigenous populations, exacerbating their rates of "incarceration and criminalization". They also point to forms of justice and punishment which fall outside of the state, including "customary forms of dispute resolution" (Carrington et al., 2015: 3). Furthermore, countries often have high levels of unemployment, struggle to reduce poverty and suffer from inequality (Carrington et al., 2015: 7-10). This recognition of the role of inequality and its criminological significance is key as it "seems almost custom-designed to produce levels of discontent and frustration that could translate into violence" (CSVR, 2008: no page number). Carrington also identify a difference in Southern 'worlds of violence', pertinent to South Africa noting a "glaring contrast" and where "[m]any countries (including South Africa…) have made progress in relation to political conflict only to continue to be dogged by high levels of criminal violence" (Carrington et al., 2015: 6-7).
Adopting a political-spatial interpretation, Watson's 'Southern Planning Theory' framework calls attention to the contexts within which African planners operate, first identifying a number of challenges for planners which relate to evident "deep and irresolvable conflict" "rampant profit-motivated land development" in very weak regulatory contexts and institutional landscapes typified by "political cultures of patronage and paternalism" (Watson, 2014: 24) as well as weak and fragmented states, communities which often do not engage in dialogue with the state but employ more conflictual practices, notions of citizenship which assume a "sense of entitlement" but not responsibility, where crime and violence are pervasive and where colonial histories "are deeply inscribed in the material fabric of the city" (Watson, 2014: 27-28). Opportunities are also evident, including rapidly changing cities with potential for innovation, youthful populations and urban residents familiar with adopting resilient practices of survival, and emerging networks of institutions working to achieve urban justice (Watson, 2014: 28).
Noting Watson's (2014) caution of avoiding the production of further theoretical binaries (2014: 23) through a Southern versus Northern set of framings, a Southern Criminology lens provides analytical space to hold together the significance of crime and housing for residents in South Africa and to make sense of the contradictory findings evidenced in this paper. These particular residents have been moved from very poor conditions (from an informal settlement known as Ocean Drive In), their histories are of dispossession and insecurity in relation to their homes and rights to live in the area, they've suffered police harassment (as illegal settlers squatting on privately-owned land) alongside managing some forms of conflict through local vigilante style 'forums'. The settlement had an exceptionally violent past primarily politically motivated which has: "left deep scars and pain in the psyche of its residents" (Sutherland and Buthelezi, 2013: 76). Their histories speak to those of many marginalised urban residents in South Africa, framed by complex histories of colonial domination, apartheid, raciallydetermined inclusions and exclusions, state-sponsored violence, extensive distrust of an overly repressive police, high levels of general crime and violence, and informal, often violent forms of dispute resolution (Breetzke, 2012). These factors occur in a context of high unemployment, particularly for black South Africans, and great inequalities in wealth, education, and health. These, alongside evident spatial inequalities (racially divided cities, dislocation from employment opportunities, fragmented transportation, unequal housing opportunities), are a product of apartheid, but also subsequent post-apartheid interventions, and they determine the urban context under examination here, as well as the perceptions of safety and crime under analysis. Furthermore, histories of racially-determined discrimination in relation to housing and urban inclusion inform and underpin the material and political rationale for South Africa's housing programme, with which this paper is concerned.
The paper now considers the significance of wider academic debates in view of its Southern Criminology approach, including the role of neighbourhood change and public housing in relation to crime, and the relevance and limitations of Crime Prevention through Environmental Design (CPTED) particularly as applied to Southern cases. Preceded by an overview of the context and methodology, these frame an analysis of empirical findings where three core arguments are developed illustrating and explaining the contradictory outcomes of improvements in safety and governance alongside rising incivility before conclusions are drawn.
2. Debating the intersections between housing and crime within a 'Southern Criminology' framework 2.1. Neighbourhood change, public housing and crime Urban transformations, including gentrification, regeneration, migration, changing population composition and relocation are shown to shape crime rates in complex ways (Kirk and Laub, 2010;Popkin et al., 2012). The role of public housing, particularly high-rise flats (noted in an American context), are associated with high crime levels and have been the recent state focus of demolitions, and relocations, into deconcentrated forms of low-rise scattered public housing (Kirk and Laub, 2010: 468). The relationship between public housing (its location, quality, inhabitants, economic realities, relative indices), processes of relocation and crime are multiple (Popkin et al., 2012), with mixed impacts potentially experienced through the migration of residents into areas which receive public housing (in whatever form) as is the case in this paper. Such selective migration "may undermine informal social control because of residential turnover and instability" (Kirk and Laub, 2010: 470) although areas receiving public housing may already have relatively high levels of crime and may be disadvantaged anyway. Popkin et al.'s work shows that urban transformation processes involving relocation from poor quality public housing, in general, leads to a decline in violent crime across case cities (2012: 154) but with more complex outcomes evident too. This work is valuable in pointing to the complexity of factors shaping public housing and its relationship to crime and the role of the state in facilitating this, but is limited by its focus on forms of housing less commonly found in cities in the global South. In particular the approach excludes informal housing, and the urban geographies of public housing location can differ depending on historical patterns of spatial exclusion, particularly of the urban poor in contexts with colonial pasts. New areas of formal housing across the global South are often green-field, and located on cities' peripheries. Furthermore, this body of work takes as its focus crime levels and types of crime whereas here questions of crime and violence are broader, situated in relation to governance and policing practices, and wider socio-cultural changes which shape contexts of urban incivility.

The limitations and relevance of CPTED for a 'Southern Criminology'
Crime Prevention through Environmental Design is an extensive field of research following the pioneering work of Jeffrey (1971) and Newman (1973) who pointed to the multifaceted relationships between urban built form and crime in the context of 1970s America as well as the ways in which urban and architectural design impacted on human behaviour. Their arguments have global appeal and influence, as well as limitations discussed below, and have been advanced further through critique (Cozens et al., 2001). CPTED's proponents point to key features (road layouts, building position, security features, property types, etc.) which enhance or retard criminal opportunities (Armitage, 2006(Armitage, , 2013. Such features are central to analyses of rehousing residents into formal state-provided units where the planning of environmental design (formal housing provision, roads, road layout, plot demarcation, street lighting, etc.) takes place, and, as is argued in this paper, shapes crime experiences. Recognising the need to incorporate social factors into analyses, scholars in this field have widened their studies to extend beyond purely physical features in shaping crime (Cozens et al., 2005), again an argument that finds support here as the role of social services, governance structures and gender prove significant. Much of this work (with a few exceptions such as Brown-Luthango et al., 2016;Brown-Luthango, 2016;De Souza and Miller, 2012;Landman and Liebermann, 2005;Kruger and Landman, 2008;Monday et al., 2013 andMeth, 2016) is however located in the global North or West, affecting the debates' capacity to test and theorise urban interventions and crime in contexts characterised by 'Southern Criminologists' (Carrington et al., 2015) in significant ways. In particular, the specific characteristics of crime in contexts of informality is considered in some of these works, and is best articulated in the UN Habitat (2011) publication "Building urban safety through slum upgrading" which argues, drawing on several cases from the global South, that material interventions are only partially successful, and real gains from upgrading require wider changes, namely governance interventions which can reduce poverty and foster participation (2011: 4). These interventions and changes form part of the analysis presented in this paper and are arguably far wider than normal CPTED responses.
However, CPTED interventions can be relevant and key issues from research which draws on ideas of CPTED in the context of South Africa illustrates the salience of analysis within a Southern Criminology perspective. The work by Landman and Liebermann (2005) along with Kruger and Landman (2008) and Landman and Kruger (2009) lead the application of CPTED to the South African context, incorporating a very particular interpretation of urban planning practice which includes a careful appreciation of the impact of post and apartheid policies on urban space, politics, participation and crime alongside a more direct application of CPTED principles. Kruger and Landman (2008) concur there may be a need to recontextualise theories and practices developed in the global North/ West if they are to be of any value in countries such as South Africa. Their paper usefully prioritises context-specific factors shaping crime (at the local scale but also the more generalised national scale) as well as 'real' responses to crime in this politically fluid context. They emphasise the particular geographies of urban poverty produced through apartheid and post-apartheid planning and consider the relationship to crime, including those of informal settlements and post-apartheid settlements. The role of poverty in shaping crime is a central argument in the paper, including the impact it has on the poor's inability to prevent victimisation because of their living conditions (2008: 78). They provide a list of key challenges shaping crime and responses to crime in South Africa, which very closely supports those identified by Carrington et al. (2015) in their wider analysis of Southern Criminology. Kruger and Landman (2008) specify: The crime situation is exceptional; extreme levels of violent crime exist; there are severe levels of poverty and inequality; the particular urban form and spatial characteristics of the South African landscape; that levels and effectiveness of policing vary, among others (2008: 79). In the context of these challenges they detail the responses to crime evident in South Africa, which vary enormously depending on wealth and consider the implementation of participatory approaches to CPTED as a potential intervention, similarly to that advocated by Landman and Liebermann (2005). In this latter paper, in contrast to methods of crime prevention which foster exclusion (gating communities as a particular example), Landman and Liebermann call for integrated approaches to reduce crime focusing on participatory methodologies to engage residents in determining the location and nature of local crime. They suggest such engagement fosters better learning and identification of crime hot spots, as well as how such incidents could be managed, namely through partnerships with the police and the municipality. Participation is critical and is encouraged parallel to physical interventions by partners such as municipal planners likely to have a positive impact: "provide street lighting, maintain the vegetation and develop the vacant land appropriately" (2005: 25). Their twin focus on integration alongside material intervention speaks directly to the political histories of exclusion and underdevelopment evident across most of South Africa's cities (in contrast to areas of privilege).
Brown-Luthango et al.'s arguments (2016) are not specifically situated within CPTED traditions but they focus explicitly on the intersections between three different housing upgrading programmes in the city of Cape Town and the consequent impacts on crime. All three cases are located in historically poor and marginalised parts of what is a deeply unequal city. The styles of upgrading consequently yield contradictory outcomes in terms of perceptions of safety and experiences of crime, with the settlement experiencing most comprehensive formalisation actually revealing the greatest anxieties around crime over time, as well as spikes in particular criminal acts, including gang related violence . Brown-Luthango (2016) explores social cohesion in the context of housing formalisation (using the case of Freedom Park, also employed in Brown-Luthango et al's paper) and points to rising crime in a context of housing upgrading associated with increases in alcohol and drug use, the dissipation of a sense of community, and growing individualism. In both papers the overriding significance of structural factors shaping residents experiences of crime (unemployment, poverty) etc. are emphasised and the key focus is on crime in relation to the upgrading of informal housing. In contrast, this paper explores the issue of relocation to new housing in a new location (similarly to Ross, 2010Ross, , 2015 exploring the opportunities offered by new-build structures, road layouts, lighting and electrification. Focusing also on the significant role of poverty, politics, and inequality, Meth (2016) argues that the design of informal housing in South Africa does indeed shape experiences of crime, as shack properties built from plastic, corrugated iron, wood and concrete, commonly suffer from 'hyper-permeability' facilitating easy access on the part of criminals. This reduces residents, in particular women's abilities to defend themselves from threats, particularly when toilets and water are located externally. Hyper-permeability applies to particular forms of informal living but not all, as those living in high-rise apartments which are still deemed informal, are less likely to suffer from permeability issues, simply because of material differences in structure. Arguably, hyper-permeability, marks the subsequent experiences of re-housed residents. The paper turns now to discuss the context of the housing programme and the case of Hammond's Farm along with the methodology underpinning this set of arguments before turning to an analysis of safety, governance and incivility in a new housing context.

Context: the national housing programme and the move to Hammond's Farm
In response to the racially-based and class-determined dire inequalities in home ownership and access during the apartheid era, the postapartheid government initiated an ambitious housing programme to counter such trends. Since this time, the Department for Human Settlements in South Africa claims to have provided around 4.3 million housing opportunities (DHS, 2016:20) as part of the substantial national subsidised housing programme with resultant housing colloquially termed RDP houses (short for the former national Reconstruc-tion and Development Programme). Beneficiaries, often previously living in shack housing, must meet particular criteria (see Tissington, 2011), as the programme largely targets the poorest in society. Housing in this policy context refers to the formal top structure, plot demarcation, roads and passage ways, infrastructure, and ownership rights. Wider neighbourhood facilities may be part of a housing intervention but this is inconsistent. Substantially reviewed (Marais, 2015;Charlton, 2013;Tissington, 2011), it is considered a success in many respects, including from the perspective of residents. It can be interpreted as a significant welfare measure (Venter et al., 2015), and celebrated for fostering citizens as 'co-sovereigns' who have heightened senses of empowerment through ownership (Gunter, 2013). The programme has however been substantially criticised for reinforcing apartheid spatial planning outcomes (Huchzermeyer, 2001;Tissington, 2011), because housing is often peripherally located, developed in sites with no facilities or economic infrastructure, as well as also falling short of gendered aims (Charlton, 2004). It has been dogged by poor administration particularly in relation to allocation practices (Rubin, 2011), and importantly, it has been criticised for failing to unseat entrenched poverty or improve livelihoods (Ross, 2015(Ross, , 2010Cross, 2013;Tissington, 2011;Charlton and Kihato, 2006). Despite these critiques, there is evidence that many residents are very grateful for their new houses and that issues around shelter, identity, and citizenship are key (see Charlton and Meth, forthcoming).
The Housing Division of the eThekwini (Durban) Metropolitan Authority has, alongside many other housing developments, recently constructed a large cluster of double-storey row houses (which residents call 'flats') in a new green-field settlement named Hammond's Farm (HF) on the outskirts of Verulam to the north of the city very near to the southern end of the new King Shaka airport. Effectively a new neighbourhood has been constructed including formal road layout, electrification, street lighting and plot demarcation. As yet no schools, community facilities or formal economic or retail spaces have been planned or built and residents rely on a pre-existing Spar supermarket complex with other smaller stores located nearby, as well as facilities in the adjacent areas of Waterloo and Verulam to sustain their needs.  (Ngcongo, 2014). This atypical housing form contrasts with the usual detached subsidy housing evident nationally, and represents more recent trends in design, aimed to maximise precious urban land through higher density building.
New residents at HF have been moved from poorer quality housing in different parts of the city, resulting in a mixing of communities and individual households from diverse locations. One key origin was that of Ocean Drive In (ODI), a severely deprived and very poor quality informal settlement situated on private land approximately 17 km further north of the city but nearer to significant, but often informal, employment opportunities afforded by wealthier Indian and white householders offering casual work (see Sutherland and Buthelezi, 2013 for details). ODI's physical qualities were dire. Houses were constructed of flimsy materials providing inadequate shelter. The settlement lacked electricity and the approximately 1400 residents depended on six communal taps, evidencing poor servicing (Braathen et al., 2014: 14) and resulting in long daily queuing for water and burdensome carrying (Sutherland and Buthelezi, 2013). Residents established informal drain pipes to manage waste water, but disconnections in their arrangement lead to water flowing into neighbouring properties, resulting in tensions between residents (Sutherland and Buthelezi, 2013: 77). Only one decent-quality road existed in the settlement, with remaining paths often narrow, unsurfaced, and subject to flooding becoming "unhealthy and un-navigable routeways" (Sutherland and Buthelezi, 2013: 76). It is residents who originated from ODI who are the focus of the study here. Their new location as they were moved into their new formal housing, adjacent to the large formerly Indian township Verulam and another older RDP estate, Waterloo, meant a localised increase in competition over scarce job opportunities, poorer access to work, and a resultant growing reliance on food handouts from local religious and governance organisations. This acceleration of poverty shapes residents' experiences, including of crime, and it parallels trends identified elsewhere as marginalised locations for new housing along with "the cost of home ownership … have also limited the effect of housing provision on poverty alleviation" (Tissington, 2011: 62).

Methodology
The empirical findings analysed here are drawn from nine focus groups (three and six with men and women only respectively, with five participants in each) and solicited diary writing combined with autophotography with eight residents (including four men) conducted in Hammond's Farm between November 2014 and June 2015. Residents were recruited through contact with local community leaders and following agreement to participate, were asked to share their experiences of moving into Hammond's Farm, changes or continuities in experiences of crime and violence (including domestic violence) and any wider social and political issues shaping their everyday lives. The use of multiple qualitative methods to explore what are often delicate social processes is central to this research, providing residents with different spaces and opportunities to express feelings and recount events in their lives which may be distressing (Meth and McClymont, 2009). This is significant when working on questions of violence and gender, and the inclusion of solicited diaries in this methodological mix, provided a particularly comfortable and productive site for participant engagement. Six key informant interviews were conducted: Area Committee member, Community Policing Forum (CPF) member, Social Worker, Police Officer, Ward Councillor's assistant, and Housing Officer. The majority of the work was conducted solo by Sibongile Buthelezi, in the dominant local language isiZulu, but Paula Meth was present for all key informant interviews and some focus group discussions. This work builds on the invaluable research conducted in Hammond's Farm by Sibongile Buthelezi and Cathy Sutherland from the University of KwaZulu Natal over the past few years for different funded projects. All transcripts were analysed using a thematic coding approach, identifying key codes and categories of findings as well as noting key absences. Analytical interpretation of data was considered by both authors and discussed with participants at dissemination events in late 2016. This paper considers some of the key themes identified through analysis, and as explained above, forms only part of the wider project's findings (which included a comparison with a case in urban India) but which also had a strong gendered focus. Here the specific focus is on changing experiences of crime and safety as a result of housing change and relocation.
5. Increased personal privacy, space and security: "formal houses have an impact on crime levels" (♀ FG6, HF, 2014 1 ) In HF, new housing is in the form of double-storey row houses with two bedrooms located on the upper floor and a private bathroom on the ground floor alongside the shared kitchen/diner/lounge area. These internal services and formalised divisions of space are significant for privacy, shaping both intra-household use of space, but also the spatial practices of visitors who now can access designated 'social' spaces (the lounge) avoiding bedrooms (♂ FG3, HF, 2014). Framing resident's perceptions of privacy, spaciousness and safety in terms of a 'Southern Criminology' foregrounds the particularities of internal living condi-1 (♀ FG6, HF, 2014): This reference denotes it was female Focus Group 6, at Hammond Farm in 2014. All participants' names have been anonymised and re-coded with different initials.

P. Meth, S. Buthelezi
Geoforum 82 (2017) 77-86 tions of shack housing as significant for understanding residents' current perceptions of safety in relation to their changed living conditions. Material qualities of informal housing across the global South are not readily generalizable but overcrowding, lack of services and unhealthy living conditions are common (UN Habitat, 2003) and reflect those of residents' former shack housing at ODI, which often lacked formal room divisions or clear social divisions of space (with sleeping, cooking, socialising often occurring in a single space). Separation of sleeping areas for parents and children was thus a key gain: "yes it is better now because even those who stay with their children have a bedroom for children. At the mjondolo 2 it was difficult having one room when you are staying with children, the mother and father" (Interview Area Committee, 2014). This issue of (the absence and then subsequent) gain in privacy from children is critical for residents and emerges as a priority issue in varying contexts (elsewhere in eThekwini and Trivandrum in India) as it directly affects adults' sexual practices (and children's sexuality) (Meth, 2015and Meth and Rajasekhar, 2016. The double-storey structure provided additional flexibility with families able to use the downstairs part of their property for sleeping (Interview Area Committee, 2014). These observations support arguments based elsewhere which evidenced the impact that cramped living conditions and loss of privacy has on experiences of intra-household tensions (including domestic violence) but also neighbourly tensions (Meth, 2015). Living in 'slum conditions' directly shapes residents' experiences of insecurity (UN Habitat, 2011) and these include social, political, governance and also physical conditions. Much of the literature on interventions in informal housing concentrates on infrastructure (see UN Habitat, 2011) yet, the material properties of informal housing directly affects residents' security too (Meth, 2016), and thus movement into formal housing directly impacts on residents' safety: LD explains how the plastic construction materials of their former shack housing enhanced vulnerability to crime: "some [criminals] were using a knife to cut the plastic [materials] of the house to get inside." (♀ FG2, HF, 2014). Residents noted the presence of locks and burglar guards as key to their new experiences of safety. Such devices are often unavailable, or compromised, in informal structures because they are too costly, or the build-quality of walls, window casements and doors, is insufficient for their use. Male and female residents commented on this different sense of safety: "I feel safe and protected here at HF; even if I go somewhere I do not feel worried about [acts of] house-breaking" (BB, ♂ diary, HF, 2015).
The formal arrangement of the housing was cited as reducing criminal opportunity and to increase police access and surveillance. At ODI theft was cited as a significant problem, including theft of washing left out to dry, but experiences of crime were often more violent, including sexual violence. Hammond's Farm's plot layout is highly regularised (see Fig. 1) (arguably monotonous) and housing construction is formally managed and allocatedall in steep contrast with ODI: "… here [at HF] … there is no place to hide… Some children were raped in the sugarcane at ODI and the person just ran away and hid there… at the mjondolo [there] was no structure people [could] build wherever they see the space.
[its] … difficult for the police to … follow the criminals" (♀ FG6, HF, 2014). Land claims processes and housing construction in informal settlements often results in organically laid out settlements, and informal governance structures controlling who builds where, often results in relatively easy access to parcels of land or housing which then impacts on the layout and organisation of the overall site. Sutherland and Buthelezi (2013) note that at ODI access to housing was historically informally obtained through a small cash payment, "a bottle of whiskey and a case of beer" (2013: 75). In contrast the regularity of layout in HF was noted positively by residents in this study, but such neatness of layout can also cause anxiety and navigational concerns (see Ross, 2010: 66) for residents accustomed to more organic "higgledy-piggledy" (Ross, 2010: 67) arrangements, furthered by confusion over the identical nature of each property (Ross, 2010: 67).
Partially a function of plot layout in this case, generic visibility plays a critical role in structuring a sense of safety for residents affected by the presence of electricity (internal to housing but significantly street lighting), the formalised housing layout, 'eyes on the street' and mobility patterns in relation to crime management. Residents described neighbours keeping watch of their homes with the double-storey structures proving advantageous as their height increases the scope of the vista from the property. The mixed impacts of electrification are discussed later, but a final particular feature of Hammonds Farm is the single entry and exit point into the settlement offering aggrieved residents an opportunity to corner suspected criminals (although arguably this might work equally effectively for the policing of HF residents). These findings explicitly support work on CPTED which points to the value of these design features in reducing opportunities for crime but which illustrate the significance of a 'Southern Criminology' lens in shaping the particularities of material changes in contexts of informality, and it supports the UN Habitat's (2011) claims about the positive role of lighting in relation to reducing vulnerability (2011: 28) in 'slum conditions'.
Residents also described improvements in their personal quality of life and safety relating to the gains in, and location of, services and building materials which compared very favourably to those at ODI: "… I am so happy to come here with my child because at [ODI] we were using izinyokanyoka (illegal electricity connection) which is not safe…. for me I have peace in my house because my old house was leaking and [it had] holes [and] rats ran inside" (PB, ♀ FG2, HF, 2014). The change from reliance on paraffin to electricity for cooking reduced the presence of smoke. Cleanliness of water alongside its presence within the home was a key gain shaping residents' well-being and safety, bearing in mind the tensions over disposing of waste water in ODI identified earlier, as well as the daily queuing and carrying of water to properties: "I am so happy to have clean water because the water we were using there [at ODI] sometimes was coming [out with] something like sewage with dirty toilet papers. We were drinking dirty water" (PZ, ♀ FG2, HF, 2014). Pit latrine toilets at ODI were described as 'dis functional' (Braathen et al., 2014: 14) and residents perceived sharing toilets at ODI as a source of disease. Significantly, the external location of toilets at ODI and the failure of existing pit latrines to operate meant residents relied on using the surrounding sugar cane for toileting (Sutherland and Buthelezi, 2013). This significantly increased the vulnerability of residents to crime and sexual violence, meaning for many women and children, going outside to the toilet at night was not feasible (Sutherland and Buthelezi, 2013), an issue corroborated through this research too.
Two new perceived dangers were however noted in relation to the materiality of new housing: namely the vulnerability caused by having upstairs windows and the dangers of an internal flight of stairs. Both were perceived by residents to be potential locations and 'tools' when fighting "It is not safe because if people are fighting the man can throw the woman through the window" (♀ FG6, HF, 2014), with this perception based on sporadic incidences of violence in HF where such practices had allegedly occurred.
Views on whether crime and violence levels were higher at ODI than at HF varied, with most claiming ODI was less safe, but not all. Sutherland and Buthelezi (2013: 76) explain that ODI has a chronic history of violence, largely tied to party political tensions between residents which resulted in extreme outbreaks at particular historical junctures. Many residents in HF come from other townships around the city of Durban and they compared HF very favourably to these townships in terms of crime levels (Interview Area Committee, 2014), this contextualises HF within a wider urban geography of comparatively high violence levels which have afflicted South African townships for decades (Breetzke, 2012). Indeed Breetzke explains (2012: 304 after Schwabe andSchurink, 2000) "that police precincts located in the townships had exceptionally high total crime counts with all violent crimes committed being above the national average".
Tied to residents' experiential knowledge of high violence levels at ODI and other townships, there's a clear sense of anxiety that the current relatively low crime levels will change once the settlement is fully occupied or has had time to 'change' or consolidate. Residents pointed to current differences in crime levels in parts of the settlement that were fully occupied (identifying for example the houses located on the higher ground/top of the settlement) where fighting and crime had grown, compared with other parts of the site which were still partunoccupied (♀ FG2, HF, 2014). Patterns of crime and perceptions and experiences of neighbourly tensions are shaped thus by temporal changes of densification, population growth and 'settlement maturity'. Here the rapid urbanisation of the city periphery, and use of green-field land to densify residential occupation of urban land is typical of cities in the global south, and raises questions of how geographies of urbanisation in such contexts tie to particular changes in crime and violence.
Resident's anxieties related also to the mixed geographical-origins of incoming residents at HF, with a large cohort coming from ODI but many others from surrounding townships "I think it is because we are still new in this area … At the moment people in this area are respecting each other" (♀ FG1, HF, 2014) and "I will say the crime will be more here once everybody is here things will change. In ten years to come you will see…" (♂ FG3, HF, 2014). These concerns about rising conflict over time parallel concerns elsewhere in relation to the consequences of housing change for community relations. Seekings et al. (2010) argue that contextual factors are significant in determining whether wholesale relocation of residents into new settlements (versus localised upgrading) have greater impacts on the likelihood of communities being effectively established. Their review of literature on this concludes that "through housing delivery the state intervenes in a complex social landscape which is characterised by a number of existing inequalities and social ties which are vulnerable to breakage" (Seekings et al., 2010: 41). Arguably, shared histories can be significant in shaping strong communal bonds (after relocation or upgrading), but this is not always the case, as these too can dissipate over time as incoming residents and changing everyday politics transforms communal relations from positive to conflictual (see Brown-Luthango, 2016 as an example).

Shifts in governance of crime and violence
Residents moving into HF experienced significant changes in governance practices, particularly relating to everyday policing and their engagement in formal layers of local governance. In addition, the temporary presence of private security in the settlement, whose role it was to protect newly-built unoccupied properties, formed an additional positive, albeit time-bound governance change experienced by residents upon moving from ODI to HF. Residents were near unanimous in their approval and commendation of the interventions and measures adopted by the then local Ward Councillor, 3 the police and local committees.
Their respect and praise for the police stands in marked contrast to low levels of satisfaction and expectation over policing and the criminal justice system identified elsewhere in the city and country more broadly (Cooper-Knock and Owen, 2015) and can be partly explained by 'good policing' alongside contrasts with shack-living which afforded very poor policing for various reasons (difficult environment for police to control, distant location, conflictual relations with police). Police made their presence felt in HF through day and night patrols (♀ FG1, HF, 2014 and♂ FG1, HF, 2014). At ODI poor access limited the police (♀ FG4, HF, 2014) and police were criticised for failing to turn up at all. In addition Cooper-Knock and Owen (2015) explain the continued relevance of the police in the South African (and Nigerian) context which despite negative understandings of their effectiveness relates to perceived roles of the police as 'regulatory authorities' and as capable of performing bureaucratic roles (the opening and closing of cases for example) (2015: 357). These regulatory and bureaucratic practices were noted by HF residents. The police were favourably contrasted with the local informal forum tasked with managing crime at ODI, which similarly to other documented informal forums engaged in 'illegal' beatings of suspected criminals. In contrast residents claimed that the police practiced formal 'proper' procedures of responding and removing suspects to the station for questioning rather than "arguing" with suspects in the neighbourhood (♀ FG5, HF, 2014). New structures of crime management and technologies of governance were established and utilised at HF providing residents with a greater sense of safety and trust. The police worked with a number of community representatives in nearby Waterloo, and used WhatsApp to communicate concerns about crime in the area (Interview Police Officer, HF, 2014). The establishment of formal community policing forums entrenched residents' confidence and underscored the value of bureaucracy, procedure and participation: "… we are safe because the police came here to elect the [community] policing forum which is working together with the police. If a person is doing a bad thing the [community] policing forum phones the police and they come quickly" (BB, ♂ diary, HF, 2015). Importantly residents' new status as 'legal citizens' also shaped their interactions with the police when previously the residents regarded the police as their enemy in response to their history of activism and protest over their poor housing conditions, now their residence in a formal 'legal' settlement bestowed residents with transformed relations with the police (Interview Police Officer, HF, 2014).
Supporting Bénit-Gbaffou observations of Ward Councillors as "local social peace keepers" (2008: 28), residents described their interactions and use of the Ward Councillor in similar terms. Residents operated within relatively formal procedures to access and utilise local committees as an initial strategy for resolving issues. If this strategy was unsuccessful, residents were advised to approach their Ward Councillor for 'peace keeping' who in turn utilised or recommended police intervention in more challenging cases. Residents supported this level and formality of governance organisation and contrasted it with ODI where "life at mjondolo is not organized" (♀ FG2, HF, 2014). The effectiveness of this hierarchical chain is evidenced through the dominance of resolutions at the local scale (♀ FG4, HF, 2014). Additionally, the more immediate location of the Ward Councillor at Waterloo, compared with ODI, was critical to this success (♀ FG3, HF, 2014) suggesting the spatial geographies of relocation worked effectively to improve political inclusion (in contrast to economic as explained earlier).

Increased 'townshipization' of everyday life
"These houses have brought different things into our lives because those who were not drinking at ODI are drinking now. People keep changing their relationships or partners. People lose respect for their partners but at ODI people were tolerating each other and not behaving like they are here. I don't know maybe people want to adopt the style of the townships to do what they want to do".SN ♀, Diary, HF, 2015 emphasis added The transition from living in shack settlements, often organically arranged, to residing in formal housing in a new neighbourhood with a mix of residents from various locations engenders significant social change (see Ross, 2010;Seekings et al., 2010). When housing delivery is large scale and on a green-field site, processes of socio-cultural change occur as the settlement and its new inhabitants 'bed down' and establish new relationships, networks and practices. Ross (2010) in the context of 'The Village' in Cape Town (a new settlement inhabited by former shack dwellers living elsewhere) describes how residents at first labelled the settlement as a "Ghost Town" so lacking in sounds, smells and sociability (2010: 68). In HF in contrast (probably a function of research timing relative to habitation) many residents noted shifts towards practices and sensory experiences akin to urban living elsewhere in the city. Labelled here as 'townshipization' this term denotes a process identified in other contexts in South Africa and is used by residents who describe the transition of their new neighbourhoods into that resembling "township life" (e.g. resident Fred, cited in Charlton and Meth, forthcoming) in a material and cultural sense. Although this process is weakly defined, residents in HF also employ the analogy to compare HF to the daily rhythms and routines evident in other black formal townships in South Africa, recognising that townships are diverse and changing spaces. Jürgens and Donaldson (2012) and Jürgens et al. (2013) review literature on changing processes in South African townships and identify various recent processes as characteristic including: demographic change and mixing, differentiation, and rising consumption, pleasure and leisure practices (2012: no page numbers). In Jürgens et al. (2013) townships are listed separately to 'RDP estates' (the latter a descriptor pertinent to HF) and in analyses of recent changing township realities they also emphasise social processes of differentiation which concomitantly is associated with rising xenophobia. Additionally ongoing "social pathologies" are evident as a result of persistent unemployment and poverty including malnutrition and violence among others. A final recent feature of townships is that of the rise of "urban revolts" in the form of mass service delivery protests (Jürgens et al., 2013: 258).
Some of these processes are certainly evident in HF and shape residents' concerns about concomitant effects including recent outbreaks of xenophobic violence, in part a function of the presence of foreign nationals in the settlement. Concerns over noise occupied many, tied to new consumption and leisure practices in HF, including drinking, playing pool and gambling and associated (rising) criminal incidences, usually fighting when drunk. In addition, changes in the way residents used public space, particularly the streets, were identified. These changed spatial practices were facilitated by lighting, specifically the presence of street lighting, decried as "police lighting" often "uniform and homogenous" (Schivelbusch, 2003: 88) but in this global South context, also which affords significant sensorial changes and opportunities for residents: "People are excited to walk on the road even at night because there are street light" (♀ FG5 HF, 2014)note the term 'excited' was articulated by several participants. This new use of the streets as a leisure space also caused concern, with many commenting on the vulnerability of children in particular. Multiple references were made to the speed and presence of vehicles witnessed in the settlement leading one representative, to reproach children who played outdoors because of anxieties about their general welfare (Interview Area committee, 2014). 'Townshipization' might also be evidenced through the labelling of the settlement's areas in accordance with locational names from central eThekwini (Durban), namely Mhlanga and Berea. Berea was categorised with reference to its busy nightlife and noise issues, its destination for incoming taxis, mirroring the busyness of the original Berea in the city centre (♀ FG6, HF, 2014) making the area popular with residents (♂ FG3, HF, 2014).
Finally, the mixing of residents from multiple source locations, including foreigners is a feature of a 'vibrant' township. In HF, mixing is directly shaped by the housing programme in different ways. First as with other housing subsidy settlements, there is a tendency to initially exclude undocumented immigrants as only South African citizens and permanent residents are eligible for housing subsidies. Overtime however as beneficiaries choose to rent out or illegally sell properties, those previously ineligible have been able to establish a home in the settlement. Additionally, the allocation process beyond residents from ODI results in a relatively random assemblage of new communities which are likely to have lower levels of social cohesion, certainly at the beginning of settlement formation. These processes of mixing caused pleasure for some, but also consternation over the social impacts of such difference and encounters with violence: "At ODI it was easy to control people by the committee but here I notice it is going to be more difficult because people are coming from different places. We also have foreigners here like Nigerians" (♂ FG3, HF, 2014). Foreigners' rental of properties from local owners provided crucial income to housing recipients but many residents blamed them for criminal incidences, including infidelities and other 'unqualified' allegations. Xenophobic violence generally caused distress to local residents who were concerned for 'foreigners' well-being but who were also anxious about personal vulnerability and the prospect of revenge crimes: "One woman came to my house and asked to hide in my house and I didn't allow her to stay in my house because I also feared that if they come and find her in my house they will also kill me with my child" (AS ♀, Diary, HF 2015).
As indicated above, concerns about the noise levels at HF related to the presence of electricity within properties, and street lighting which fostered greater use of street space as a leisure space through the night as well as the opportunity to play music throughout the day and night from homes (♀ FG6, HF, 2014). The presence of light through electrification was charged with influencing levels of fighting, as "people are not sleeping and are drinking until the morning because it is light" (LN ♂, Diary, HF, 2015). Noise negatively impacted on families and also those who had to get up and go to work the following morning (BB ♂, Diary, HF, 2015). The association of "incivility and crime" within urban spaces at night (and the subsequent policing and control of such practices) (Crawford and Flint, 2009: 405-406) is a common attribute of cities (albeit with varying features), whereby "the night-time economy constitutes a place in which disorder is an essential by-product of a brand of alcohol-infused consumption" (Crawford and Flint, 2009: 407). This issue of alcohol consumption is returned to below.

Snooker, gambling and fast cars
Perhaps symptomatic of township life, men in particular identified snooker and gambling as two positive additions to their leisure experiences claiming they helped keep people out of trouble, through reducing levels of boredom. Male residents expressed real delight and gratitude over the introduction of snooker to their settlement by fellow entrepreneurial male residents and also identified how such leisure practices fostered engagement with other members of the community explaining "Everybody goes there and it is always full" (MG ♂, Diary, HF, 2015) These new leisure practices brought danger however with allegations of force and violence identified in relation to gambling (♀ FG6, HF, 2014). The issue of fast cars identified above, indicative of rising consumption, was also a persistent complaint, with some identifying foreigners as rogue drivers. Many residents called for the introduction of speed bumps into the settlement's roads in order to manage speeding. This concern and proposition is indicative of formalisation, hitherto impractical at ODI where roads were unsurfaced, but also unnecessary at ODI, since the informal settlement lacked many roads navigable by vehicles.

Rising sales, use and abuse of alcohol
As discussed above, the noise made by residents was frequently linked to drinking habits and fighting. The location of the large Spar supermarket at the top of the settlement, in very close proximity to much of the housing, and its sale of relatively cheap alcohol until 8pm was noted as central to the rise in drinking habits and the ease of accessing alcohol. In addition, residents' new-found possession of refrigerators facilitated cool storage (♂ FG3, HF, 2014). Drinking increased over the weekends and also at month end, after salaries were paid: "It is difficult to sleep while people are fighting on the road, singing, walking up and down the road and making a noise" (JB ♀, Diary, HF, 2015) Residents claimed peace and quiet was only obtainable through police intervention. Various shebeens 4 and houses selling alcohol (many which have no closing times) have opened up in HF, selling alcohol purchased from the Spar allowing what was perceived as higher rates of alcohol consumption than previously experienced at ODI (♀ FG5, HF, 2014). Such rising trends in the illegal sale of alcohol directly parallel findings in Cape Town (Brown-Luthango, 2016;Brown-Luthango et al., 2016). The practice of drinking alcohol facilitated social mixing, a process central to an idea of 'townshipization' and which fosters the merging of residents from differing backgrounds viewed as a positive outcome by some residents: "… people are excited and we are mixing with different people from Inanda, Ntuzuma [Mlazi] and Phoenix. If you can come here on a Friday you will see what is happening. Girls and boys are the same" (♂ FG3, HF, 2014).
High alcohol consumption was confirmed by the local police department who explained that it wasn't homemade beer being sold but rather 'Castle Lager', and spirits such as cane being sold. Police responded to these illegal shebeens by searching for and charging such practices with a view to their closure, on an average of around 5 to 6 per month (Interview Police Officer, HF, 2014). Links between violence and alcohol consumption were argued, by the police and community representatives to be significant: "The only thing which will destroy this area is alcohol because [consumption] is very high" (Interview CPF, HF, 2014). These concerns included impacts on domestic violence where 80% of incidents were seen by the police as tied to alcohol (Interview Police Officer, HF, 2014). Alcohol was only part of the story however, with the Community Policing Forum's statement below suggesting that additionally residents emanating from ODI were fundamentally violent (supporting evidence above about residents from ODI suffering and engaging in extreme violence historically) but locating this violence as integral to informal living: "The men were fighting and they stabbed each other because of alcohol. The people who are coming from Ocean Drive In are so violent … Last week again people were stabbing each other at Platform 8. 5 What causes all of this are the people who are selling alcohol. The people who are coming from informal settlement are still behaving like as if they are still living in the shacks".Interview CPF, HF, 2014 Comparatively speaking however, HF was considered to be generally peaceful and complaints about assaults for example occurred mainly over the weekend when people drank more. The police claimed that people were more receptive to being policed now they were living in HF (Interview with Police Officer, HF, 2014). Drinking and fighting is also place and time specific within HF, with some residents claiming to not witness fighting and aggression between drunk individuals (FG1, HF). Finally the abuse of alcohol and drugs was justified as a solution to boredom, itself a likely function of poverty, unemployment and the lack of alternative leisure facilities, and educational facilities in the settlement as the extract from TN illustrates: "The level of using drugs here is increasing … Even people who were not drinking at ODI are drinking here now… I am saying that because I'm the one of those who was not drinking but now I am drinking … I noticed that my life is not moving forward because whatever I get I spend on drinking alcohol. … For me if I'm drunk I always want to fight with someone. I do not like this behaviour but I'm trying to change…. The youth are bored, that is why they are drinking too much. Nothing keeps the youth busy".TN ♂, Diary HF, 2015

Conclusions
Using the lens of Southern Criminology (after Carrington et al., 2015) this paper has analysed the changes in experiences of safety, crime, violence, everyday incivility and some governance practices as residents have moved from an informal settlement to one consisting of new state-subsidised formal housing. The resultant trends are complex and at times contradictory and also spatially and temporally specific, with the context of urban transformation in post-apartheid South Africa key to the interpretation of such findings. The new housing brings significant benefits to residents in terms of privacy, security and day-today safety of living in formal housing, supporting elements of CPTED principles of design employed to reduce crime, but recognising sociopolitical practices remain key (Cozens et al., 2005). Clearly shifts away from the 'hyper-permeability' (Meth, 2016) of slum housing is significant but not a sole guarantor of safety, but this hyper-permeability is a critical feature of Southern Criminology and sets apart the particular vulnerabilities of residents living in such material and political conditions. Recognising the context of a history of dispossession which has fostered racism, exclusion and discrimination, shifting political landscapes (along with the criminal justice system) and their relationship to experiences of crime is central to understanding residents' attitudes and responses to governance mechanisms. Hence, at times residents embrace the reach and power of the state, at multiple scales, to provide security and protection. This included a desire to support the bureaucracy and formalities of policing alongside their willingness to engage in formal participatory structures at the local scale. Their 'tentative' willing is tied to their histories of exclusion and their adversarial engagements with the state when living informally. This reveals the potential embedded in the varied and often positive role of governance institutions in mediating safety. Finally, the 'townshipization' of the settlement brings with it a host of material and socio-cultural changes, namely electrification, community mixing, including with 'foreigners', alcohol consumption, fighting and other leisure pursuits. In contexts of high unemployment and blatant boredom, particularly for younger residents, these tie directly to experiences of crime and everyday incivility coupled with ongoing processes of urban consolidation and neighbourhood change. Fears of future vulnerability underscore the temporal nature of neighbourhood change and perceptions of crime, again a feature of rapidly transforming urban spaces typical of many cities in the global South. In this regard the findings at HF mirror those identified by Brown-Luthango et al. (2016) although this case is distinct in its experience of relocation rather than in situ upgrading. Furthermore, this work supports the challenges of new peripherally-located housing identified by other authors (Cross, 2013;Tissington, 2011, etc.) but in fact residents here were housed closer to the city's economic heartland, but paradoxically their employment opportunities were worsened, for very contextually-specific reasons tied to particular historically-informed geographies of racial settlement (Indian, white and African locations), dominant private sector land ownership and development, and poor transport connections which limit opportunities to climb out of poverty. Reducing inequality and poverty are critical to maximise the housing asset in ways which could be cumulatively beneficial for men and women including in terms of safety. In short, crime can be reduced through the provision of state-subsidised housing for former shack dwellers, but such gains are temporary and explicitly connected to and dependent upon wider economic trends which can work to entrench and worsen some of the challenge of such cities in a "Southern" context which mark crime and governance in such particular ways.