Introduction

Towards the end of every month, hundreds of handwritten advertisements appear on a wall next to the main supermarket in Yeoville, Johannesburg (see Fig. 1). Almost all of the notices advertise ‘rooms’ and ‘spaces’ available for rent. These refer to the portion of a flat or house that is let as a personal living space, where occupants share facilities and common areas.

Fig. 1
figure 1

The Yeoville Community Board wall at the end of the month, showing hundreds of advertisements, mostly for rooms and spaces for rent. Lessors await potential lessees, and bakkies (vans) park alongside, ready for hire to move peoples’ belongings (source: author’s own photograph, November 2012)

Full size image

Rooms and spaces are a phenomenon reported across the Global South, although often referred to by different names and exhibiting varying characteristics and contexts. These often take the form of single rented rooms such as documented in Nairobi (Andreasen 1989), in compounds of single storey rooms arranged around a courtyard, or in a multi-story variant in Ghana (see for example Korboe 1992; also see Tipple et al. 1994). Other documented examples include those found across various settlement typologies in Lagos, Nigeria (Aina 1990), in Chitungwiza (Schlyter 2003; Dewar 1997) and in Gweru in Zimbabwe (Grant 1996, 2007), Mwanza City in Tanzania (Cadstedt 2010); Mombasa (Macoloo 1991), as well as the medium and high-rise examples described in Nairobi by Huchzermeyer (2011). Rooms and spaces in these contexts are recognised to house significant numbers of people.

In South Africa too, rooms and spaces are a common feature of the housing landscape, with the version known as backyard dwellings becoming more and more common in townships (Lemanski 2009; Shapurjee and Charlton 2013). In inner-city Johannesburg,Footnote 1 rooms and spaces are now widespread in this district of dense, multi-storey and predominantly high rise buildings, although in most cases they have emerged without the approval of authorities. Whilst these rooms and spaces are alluded to in several studies, including Carey (2009), Dörmann and Mkhabela (forthcoming), Few et al. (2004), Gordon and Nell (2006), Hellman (1935), Lund et al. (2004), Poulsen (2010) and Tissington (2013), specific information is lacking: the typologies are referred to by varying names, or go unnamed, and are often spoken of as undifferentiated from the surrounding unit in which they are located. There is little systematic analysis of the phenomenon, understanding of its characteristics or the reasons for its prevalence, and little documentation of the perceptions of its occupants.

At a national level, South Africa is faced by what is often referred to as a ‘housing crisis’ (e.g., Community Law Centre (CLC) and SERI 2013). Section 24 of Chap. 2 of the Constitution established in 1996 that ‘everyone has a right to adequate housing’, and there is a ‘plethora’ (Tissington 2013: 31) of complex policies and programmes. However, a ‘backlog’ of more than 2.4 million houses remains at a national level (Ross et al. 2010), and there is a large unmet demand for low-income rental housing, particularly in inner-city Johannesburg (Tissington 2013). National policy documents promote affordable inner-city housing (DHS 2004: 14), but a variety of legislation and policy as well as general government sentiment does not favour rooms and spaces (see below). Formal providers of housing, while recognising the value of rooms, are in effect encouraged to focus on other typologies that have more viable funding options.

As highlighted above, rooms and spaces are widespread in the inner city, much of the accommodation occurring in the informal market, despite remaining unsanctioned by the municipality. Given the extent of the phenomenon yet the lack of specific knowledge about it, coupled with the severe consequences for residents of actions taken to eliminate it, this article explores what this type of accommodation offers those who occupy it, and what can be learnt from this that can be used to inform policy responses.

The article draws on fieldwork conducted in the inner city of Johannesburg between 2011 and 2012, which investigated three buildings offering various kinds of rooms and spaces. Twenty in-depth interviews with residents were conducted across the three buildings, supplemented by participant observation by the primary author of this article. This included the lead author living for a period of time in two of the buildings. Conceptually, the discussion takes a livelihoods perspective, exploring what the physical characteristics, management practices and occupancy strategies associated with rooms and spaces offer to those who make use of them. The article considers whether the typologies act as an asset, allowing people to respond more effectively to opportunities and to be more resilient to negative changes, or whether the typologies in fact add to vulnerability.

We conclude that the diversity of rooms and spaces across the three buildings, and particularly within one of the buildings, offered degrees of flexibility and diversity that was invaluable to occupants with insecure livelihoods and incomes. Particularly significant to peoples’ ability to quickly adapt their accommodation to changes in their personal circumstances was the decentralised management, multiple levels of subletting and subdivision, and month-to-month tenancy system present in one of the buildings studied, which went hand in hand with well-suited forms of advertising, accessibility of information and an infrastructure to support relocations. This represented a more multi-faceted logistical system around accessing accommodation than described in other studies. Whilst occupants are shown to be extraordinarily resilient, findings from the three different ‘forms’ of rooms and spaces covered in the current research demonstrate an adverse relationship between accommodation and livelihoods, where the form available to people with the least secure livelihoods is that which also subjects them to the greatest insecurity (see also Schlyter 2003). While there is scope for external intervention such as government regulation to limit negative impacts, the many ways that certain aspects of rooms and spaces support occupants, given the reality of their livelihoods, need to be acknowledged and retained as much as possible.

We first outline the livelihoods perspective, then the methods used in the current study and the three case studies investigated, before reporting the key findings from the study, holding a brief discussion around implications and concluding with recommendations.

Conceptual Framework: a Livelihoods Perspective

We draw primarily on the concept of livelihoods as a means of analysing what this accommodation offers its occupants, utilising the conceptual framework used in some analyses of South African government subsidy houses (see for example Rust et al. 2009; Baumann 2003). Although the physical form and typology of the current study differs from the low-rise, detached subsidy housing, exploring the value of the housing to the occupant is a shared concern.

Whilst livelihoods refer to ‘the means of gaining a living’ (Chambers 1995:174), the livelihoods of poorer people generally draw on ‘multiple activities and sources of food, income and security’. A sustainable livelihood can satisfy basic needs and is also able to withstand ‘shocks and stresses’ (Chambers 1995:175). A related concept is that of vulnerability, defined by Moser (1998:3) as ‘insecurity and sensitivity’ to changes in circumstances, as well as peoples’ responsiveness and resilience’.

The means of this ‘responsiveness’ and ‘resilience’ are through assets (Moser 1998:3), which can take various forms. There is no a priori ranking of assets: certain assets are more valued by some people while other people value them less (Baumann 2003). Rakodi (2002:10) argues that focusing on assets, that is, an analysis of strengths as opposed to needs is ‘more empirically sound, conceptually appropriate and of more practical use’.

Importantly, it is not only the amount of assets that individuals, households or communities have but also how well they can be mobilised and managed. Moser (1998:5) describes the poor as managing ‘complex asset portfolios’, which people adjust to achieve the livelihood outcomes most suited to them. They employ both ‘income raising strategies’ aimed at acquiring and improving access to resources, and ‘consumption modifying strategies’ aimed at restraining the depletion of resources (Devereux 1993:57, as cited in Moser 1998:5; also see Dercon 2000).

However, some authors contest just how much decision-making power poor people have: they are often faced with very few alternatives and lack control over their assets, and can merely act opportunistically (Rakodi 1995). Furthermore, it is argued that policymakers and planners cannot simply rely on the initiatives of poor people (Beall and Kanji 1999), but should enable greater access to a sustainable livelihood (Chambers 1995). Importantly, however, this assistance should come in the form of promoting opportunities: supporting the assets that poor people are already utilising successfully (DFID 1999). Before this can happen, however, there needs to be an in-depth understanding of the strategies people are already having success with. The article explores how occupants utilise some aspects of rooms and spaces as assets and navigate those aspects that can exacerbate vulnerabilities. Through a livelihoods analysis, the authors are able to present some recommendations regarding the different forms of rooms and spaces studied.

Method

The fieldwork employed an ethnographic method that was centred on participant observation and qualitative interviews to develop deep insight into the way accommodation functions and is used in three different buildings. The buildings were selected from observation and preliminary discussions with residents, while input from professionals in the housing sector confirmed that the buildings were significantly different from each other and therefore spanned some of the variation in rooms and spaces found in the inner city. Selection was also influenced by whether the management or initial participants were receptive to the research.

All three buildings were multilevel blocks with over 40 room units, although one block, Hillview, had significantly more rooms than this. Aside from this the buildings were quite different from each other, including the formal purposing of the building as well as the nature of the internal partitions. The condition or level of maintenance of the buildings also varied. The key physical and institutional characteristics of each building are described below, along with how the primary author made contact and gained access.

Building 1: DonkelaFootnote 2

Donkela is a warehouse in Doornfontein on the eastern edge of the Johannesburg CBD. It was not designed or intended for residential living, but at the time of the research, it was occupied by over 200 people living in rooms informally crafted out of the interior space. The large open spaces of the warehouse meant that rooms were assembled side-by-side, forming corridors connecting the rooms to the main entrance with the stairwell and the back entrances. One corridor leading to the old office was bricked off and had a security gate separating the rooms along it from the rest of the building, which meant this cluster of rooms formed something of a unit (Fig. 2).

Fig. 2
figure 2

Corridor running between rooms on Donkela’s ground floor. The light is from the camera flash—these were mostly quite dim though areas were lit from overhead lights in the rooms (source: author’s own photograph, September 2012)

Full size image

Partitioned rooms had usually been constructed by the first occupants who had purchased the space (usually from the building committee or through a corrupt member outside of committee processes), and the materials remained there even after that occupant moved on. Most of the brick rooms had been constructed by the original ‘hijackers’ of the building. The lead author did not observe many smaller space divisions within these rooms. Unlike the other two buildings Hillview and Abney (below) where the rooms were sealed units, the walls of most spaces in Donkela did not reach the roof and were made from temporary, more porous materials. While Abney and Hillview usually had a reliable supply of water and electricity, Donkela did not. It became clear during the research that some people struggled with the very basic living conditions more than others did. Men in general seemed to find the living conditions in Donkela a lot easier than women.

Donkela’s ‘official’ management was by a residents’ committee, made up of five members who collected rent; recruited, managed and paid the cleaner and security guards; kept paper records of some financial transactions; liaised with external parties including the City, lawyers, researchers and the police; sorted out disputes between occupants and sometimes those with external parties on behalf of occupants; held meetings with all occupants to brief them about the evictions case they faced or encourage them to clean the building. In general, they were the ‘go to’ people for residents, who held knowledge of the history and current state of the building, also in relation to the surrounding environment. The organisation Médecins Sans Frontières (MSF) had helped the residents committee set up a bank account where occupants deposited the levy for their room using their room number as a reference for transparency and accountability.

However, at the same time, those occupants of the building originally from Lesotho had set up their own leadership structure, made up of men. Although ‘unofficial’, this structure performed some of the functions that the residents committee performed, including dispute resolution, liaison with external parties (on behalf of the Sotho ‘community’), and being the general point of contact for most of the Basotho. The two structures were not in direct competition, as the Sotho structure did not attempt to collect rent; however, some people from Lesotho did stop paying the ‘official’ levy, with the vocal backing of the Sotho leadership. Particularly after the chairperson of the residents committee died, many occupants had less faith in the real strength or effectiveness of the residents committee and the Sotho leadership gained prominence. Interviewee I1 (Sep 2012), interviewed after the chairperson’s death, said, ‘The so-called committee don’t work. If you go to them they say, “ah we’ll talk tomorrow”…They are scared of the Sothos…The Sothos rule us, [yet] everything they [the Sotho leadership] do is for the Sothos’.

The lead author negotiated with the residents’ committee of Donkela who indicated that he could share a room with the chair of the committee, as this room was large and secure. The lead author stayed there for less than a month, however, as he developed a bad allergic reaction to the bed bugs which were rife in the building.

Building 2: Hillview Mansions

Hillview Mansions is a five-storey block of flats in the predominantly residential suburb of Yeoville, inner city Johannesburg. The building consists of a diversity of formal units, from former ‘servants’ rooms to five-bedroom flats. Externally, each unit remains largely unchanged from its original design, but within these units, internal configurations have changed considerably, generally being divided into much smaller units of occupation than before. There are now varying sizes of spaces available for rent, at varying costs, security, levels of maintenance, spatial configurations and combinations of occupants. Although the basic level of services seemed to be similar across the building, owners, omastanda (discussed below) and/or other occupants have arranged the rooms and spaces slightly differently in each flat. Some omastanda maintained a communal living area with a TV and sofas, but sectioned off an area within the room for someone to rent. Some chose to keep a separate living area for themselves, with the kitchen, bathroom and/or toilet the only communal space. Some used hardboard to create subrooms to rent at a higher price than ‘spaces’, while others simply let out the space and occupants used their own material to string up curtains for themselves. In other flats, wall units or wardrobes were used to partition the room (Fig. 3).

Fig. 3
figure 3

The entrance of a one-bedroom flat in Hillview Mansions, Yeoville, featuring the creation of ‘spaces’ through the use of curtains in what was originally the entrance hall and living room. Five units were thus created within the flat (source: author’s own photograph, July 2011)

Full size image

The management system involved different tiers and a combination of ‘official’ management structures, competing structures and ‘unofficial’ structures. As a sectional title building, a body corporate had the responsibility of the building as a whole, with convoluted and at times apparently corrupt official and unofficial structures. However, interviewees were not aware of the dynamics of management at building level, and many did not perceive themselves to be directly affected by it, thus the discussion here focuses on the mastanda system operating at the level of individual flats.

A mastanda (plural omastanda) literally means ‘the one who owns the stand’ in the isiNdebele/isiZulu lingua franca spoken, but in this context refers to a specific type of manager (Hungwe 2013). This manager does not have to be the owner of a flat but is mostly the main leaseholder or the one that is in direct liaison with the owner, and is usually nominated by him or her. The mastanda advertises for and brings in sublettors, collects the required total rent for the flat, negotiates and regulates the use and payment of services with all occupants, assists in ensuring the flat was clean and maintained as much as possible; attends to occupants’ complaints, and manages conflict (although sometimes omastanda are the cause of it). They were almost always live-in, aside from the occasional case where the owner or main leaseholder maintained their role as mastanda even when they had moved out. Omastanda often discounted their own rental through charging the other occupants proportionally more to cover the difference, while a few managed to cover all of their own accommodation costs, and one or two managed to make money.

The lead author found the room in which he stayed at Hillview Mansions by responding to an advertisement placed by the mastanda on the wall outside Shoprite in Yeoville (see Introduction and below). The lead author stayed there for over 4 months, building strong relationships with the caretaker and chair of the board of trustees and good friendships with many of the other occupants.

Building 3: Abney

Abney is a building in Hillbrow that is run formally as low-income rental accommodation in compliance with city regulations. Rooms faced either onto (or over) the courtyard or to the exterior of the building, connected by a corridor with rooms opening off either side. All rooms were modified according to building regulations as part of the conversion process, but differed slightly in size and access to a balcony. Some occupants had created spaces by using wardrobes or shelving, but most people who shared did not partition the room further. The rooms and spaces were thus very uniform. The building provides shared ablution and cooking facilities.

Abney’s management system was more formal than the other two buildings and more in line with the organisation’s official structure. Occupants were usually in direct contact with the building manager, who lived in a room in the building with his wife and child. If there was an issue with tenants he could not handle, he escalated it to head office, usually the client relations department. There were also floor committees made up of residents, but these were not functioning during the period of interviews, though they became more prominent in the rent boycott that occurred in the period after the lead author’s fieldwork. Abney’s management company did not grant him access to stay at the building and perform participant observation. Interaction with building residents was limited to interviews.

The lead author conducted seven in-depth interviews in Hillview, seven in Donkela and six in Abney. Of the total of 20 interviews with residents, 17 were conducted in the room or space of the interviewee allowing for further observation. Selection of participants was through a mixture of criterion-based sampling and convenience sampling, due to difficulty gaining access to all occupants (see Maree 2012).

The lead author was able to have repeat interviews with select key informants from Donkela and Hillview Mansions, providing a revealing longitudinal perspective (Ritchie and Lewis 2003:54). By capturing the occupants’ descriptions and experiences after witnessing the rooms/spaces they were speaking of, the lead author was able to draw direct comparisons with his own experiences and past and current observations.

Results

Below, we discuss findings on the relationship between livelihoods and the rooms and spaces studied in these buildings in Johannesburg’s inner city. We organise the results into three core themes: location and affordability, flexibility (but with potential insecurity), and the relative importance of different aspects of the accommodation (services, safety, health, and space and privacy). Within each theme, we look at each of the buildings: Donkela, Hillview Mansions and Abney.

Location and Affordability

First, we discuss why participants in this study were living in central Johannesburg and in the particular accommodation they had found. In this aspect, the current study correlated well with the literature. Participants reported that they lived in the inner city either because they were close to livelihood opportunities or because they could easily access transport to opportunities. Echoing many other studies (see for example Custers 2001; Brown and Lloyd-Jones 2002), including those focused on Johannesburg (Mkhabela and Dörmann under review; Few et al. 2004; Poulsen 2010; Carey 2009; Gordon and Nell 2006; Trafalgar 2006), being based in inner-city Johannesburg allowed participants to reduce transport costs, to be more flexibly and reliably connected to the greater urban conurbation, or be more intimately connected with the loci of production. Living in the inner city also provided other support which allowed participants to get by more easily by cutting costs, ensuring an education for their children or building social capital. Respondents in all the buildings emphasised that ‘everything they need’ is nearby.

For some, this could be seen as an act of ‘asset management’ (Moser 1998:1), where participantsFootnote 3 sacrificed better housing conditions at home (their place of origin), potentially exposing themselves to greater vulnerability, for the employment and/or livelihood opportunities in the city. However, in this study, their real decision-making power (and thus ‘management’ potential) seemed minimal: with no perceived livelihood opportunities at ‘home’, the only real decision was which city to try. Respondent I13 (18 August 12), for example, indicated that if necessary, he would ‘go back to [the town of] Mafikeng, but job is scarce, unless you have a connection’.

For many respondents in this research, the reason they chose to stay in their particular room or space was because it was the only option they could afford. Rooms and spaces on the informal market are reportedly often the first step for many migrants seeking entry into the ‘urban system’ (Grant 1995:359), as the most affordable and flexible accommodation type (Grant 2007). In this study, a number of people in the building with the most basic accommodation and facilities, Donkela, could not easily think of alternative accommodation in the event that they were forced to leave the warehouse, giving responses such as ‘it is going to be very hard—there is no alternative place at this time’ (interviewee I4, 23 September 12), or suggesting that they would become homeless and have to live ‘under the bridge’ (I6 4 October 12). This accords with the view that people live under such poor conditions such as those in ‘bad’ buildings because of their ‘desperation for housing in the inner city’ COHRE (2005:42) and because they have ‘no other option’ (Carey 2009:19), recalling Rakodi’s (1995) concern that such circumstances may not enable much choice. Beyond affordability constraints, studies indicate that the admission policies of some formal accommodation management companies may exclude potential residents: for instance, one landlord interviewed by Greenburg and Polzer (2008:11) said ‘we’re a South African company and we have to look after our own first…it’s illegal to rent to a refugee’.

The research showed that many peoples’ incomes were unstable and fluctuated, with recognisable commonalities within the same building and differences between occupants of the different buildings. The livelihood activities of occupants of the Donkela building were particularly insecure, with many incomes sought in the informal sector. Many residents were either seeking a job or more stable income, were informally employed (e.g., selling cooked meals), involved in casual work (such as in a shop), or in formal contract/temporary work for labour brokers (e.g., offloading trucks). A few were formally employed (e.g., as a security guard) with more stable incomes.

By comparison, most Hillview occupants had generally obtained sources of higher income, or lived with partners, family or friends who had such incomes. However, this often included a mix of formal and informal livelihoods: many people’s livelihoods remained unstable. By and large, this picture resonated with other studies reporting a range of insecure incomes in the poor living conditions investigated in Johannesburg (COHRE 2005; Few et al. 2004), and/ or a low incidence of formal employment (Aina 1990).

In the third building, Abney, many occupants had found formal jobs that provided a stable income or were supported by friends or family with formal jobs who stood security for their lease or contributed to the deposit and/or lease. Nevertheless, there were some occupants who managed to access accommodation in the building through subtenanting from the main leaseholder who had qualified to sign for the formal lease.

However, the research showed that not all occupants were staying where they were because it was the only accommodation they could afford. Other respondents in the research had made a choice to minimise their accommodation costs, preferring to allocate their discretionary spending to other priorities. Respondent I5 minimised his accommodation costs by living in Donkela in order to cover the costs of his extended family. He supported his two school-going brothers and father in Mozambique, as well as his wife who lives with him, all from money he earned from selling accessories for mobile phones and other small plastic goods on the street. In another example, one of the rooms in Donkela occupied by five men who said they were living as a ‘family’. Three were working in formal jobs, and they could have afforded to rent a shared room or spaces in buildings with better conditions, such as Hillview. However, they chose to make sacrifices regarding their accommodation in order to spend on alternatives. One sent most of his money ‘home’ to support his wife and two children. Another was younger and had no dependents, saving most of his money. His other discretionary spending went on clothes and a flashy mobile phone. He saw himself as being successful in the future where he would afford a fancy apartment, but said ‘before the sun can shine it has to be dark—this [living in Donkela] is the dark side’ (I7 4 October 12).

Thus, for some participants, staying in rooms or spaces seemed an example of asset management: although they could afford accommodation with better conditions, they chose to reduce the cost of accommodation in Johannesburg’s inner city and divert their discretionary income to things they deemed as more important, including investing back ‘home’ (Poulsen 2010; UN-Habitat 2003). Literature confirms a strong connection between the rooms and spaces typologies and migrancy, both within and between countries (see, for instance, Few et al. 2004; Greenberg and Polzer 2008; Andreasen 1989; Poulsen 2010; Grant 2007). Studies show most participants to have limited interest in settling in the locale they were currently resident in (Andreasen 1989; IRIN 2011), and various South African studies show patterns of circulatory migration to be strong (Dewar 1997; Greenburg and Polzer 2008; Mabin 1990; UN-Habitat 2003). These findings challenge housing policies that advocate providing title deeds and that encourage incremental upgrading of rooms and spaces with the resident occupants (Andreasen 1989; Wang and Li 2006). Nevertheless, the direction of causality is unclear: it is uncertain, if the accommodation improved, whether occupants of rooms and spaces would prefer to settle in the city rather than migrate (Andreasen 1989; Potts 2000).

Perhaps most importantly on the matter of affordability, for Hillview occupants, the subletting of rooms or spaces offered additional income, utilised mainly as a way to subsidise their own costs of accommodation. This could be seen as being in line with Dercon’s (2000) identification of ‘income raising strategies’ to overcome livelihood shocks and trends. In many units where more flexibility was allowed, occupants were able to sublet a portion of their accommodation in times of additional financial constraints: if they were renting a room, for instance, they could sublet a portion of it as a space, or even a portion of their space, as long as they continued to cover the full amount required by the mastanda. During this research, the lead author was able to halve the cost of his accommodation by subletting the room, while maintaining a degree of privacy for himself by sleeping on the closed balcony. When respondent I9 needed to send more money ‘home’ to Kampala for one of her children who became ill, she advertised and found respondent P4 to share her space, also halving the proportion of rent she had to pay.

Abney also offered the possibility of sharing the room, with some leaseholders bringing in an additional roommate to share the rent, though the flexibility was limited as not more than two adults were allowed in each room. The practice of subletting did not seem as common in Donkela, with no participants reporting sharing for financial reasons. This was possibly because the monthly costs of Donkela were publically known to consist only of the R100 levy (no rental amount). However, financial returns in return for sharing could have occurred in a few rooms without the committee members or the researcher knowing about it.

To conclude on the matter of location and affordability, the study largely concurred with those that point to rooms and spaces being the only option affordable to many occupants, or for some such as foreign migrants, their only means of accessing accommodation. Importantly, however, some respondents in this study had chosen to minimise their living costs in favour of other priorities. Lastly, occupants were able to sublet rooms or spaces to ease livelihood shocks or trends such as loss of income or additional expenses.

Flexibility (but with Potential Insecurity)

The second theme from our findings explores how the three buildings offered differing degrees of flexibility, with a potential flipside of added insecurity for occupants in some cases.

Diversity, Mobility and Systems of Exchange

Studies have found a high degree of mobility amongst residents of rooms and spaces (Grant 2007; Andreasen 1989; Few et al. 2004 Greenburg and Polzner 2008).Footnote 4 Mkhabela and Dörmann (under review), for example, describe rooms in Yeoville as ‘sites of circulation’. Authors view this movement in different lights: some positive, some negative. Some note that occupants have utilised flexibility within the system to move easily in more financially or otherwise challenging times, within the same city or elsewhere, including back home. Other occupants were able to move as soon as a better livelihood opportunity or an accommodation option presented itself. Other authors relate mobility to poverty and lack of choice, seeing these movements as enforced. Thus, mobility could be seen as a facet of rooms and spaces that suits peoples’ livelihood strategies and enables better access to changing livelihood opportunities, or it could be seen as a factor adding to occupants’ vulnerability.Footnote 5

In this study, the typology of rooms and spaces appeared to facilitate mobility in ways that was overall positive. Advantage was derived from the variety of spaces available, and the relative ease with people could change circumstances: change location or living conditions within the inner city, or within a building, or even within a unit. There was a high degree of diversity in accommodation size and type between buildings and in the case of Hillview, even within the same building, where options were further diversified through the wide array of informal conversions. Subletting of these rooms and spaces opened up a myriad of accommodation forms for occupants to choose from.

In the same way that occupants used subletting to make their accommodation more affordable, so they used it to deal with unstable livelihoods. This was particularly prevalent at Hillview Mansions where some adapted their accommodation expenses to variations in their cash flow or to livelihood shocks they had experienced. These occupants sublet portions of their room or space to bring in income, while others would move to smaller or less satisfactory but cheaper accommodation. This offered occupants the possibility of ‘consumption smoothing’ to mitigate vulnerability caused by shocks that translated into the erosion of their financial assets or flows (see Dercon 2000). Nevertheless, other Hillview occupants, particularly primary leaseholders and those with more stable incomes, had rented the same room for many years. Thus, rooms and spaces were shown to also have the potential to provide very stable tenure.

The system of exchange of rooms and spaces in buildings like Hillview Mansions was found to be very dynamic. The wall of advertisements (Fig. 1) was a key part of it, but there were several other important elements to facilitating smooth mobility. At month-end, lessors seeking lessees waited around the wall and approached potential lessees directly. Bakkie drivers parked alongside, servicing those who had found a place and wished to utilise removal services. Taxis parked opposite, servicing accommodation-seekers wishing to view the accommodation once they had made contact with the lessor. A few additional notices advertised second-hand furniture and appliances. Sometimes people displayed household items further down the street, or the bakkie drivers dumped furniture in bad repair or no longer wanted by those they assisted to move, and other people picked these up. Furthermore, many moves within a building were facilitated by word of mouth, notices in the window of a flat or at the building entrance, going door to door or through a list kept by the security guards. Systems such as these facilitated a rapid and streamlined exchange of rooms and spaces in Yeoville (and is apparent in other inner city areas such as Berea and Hillbrow), with occupants able to complete the process of accessing the market or moving from one room/space to another in a matter of hours. Thus, one of the elements setting the Hillview Mansions system of exchange apart from rooms in Donkela and Abney, or from the predominantly word-of-month or going door-to-door systems found in the literature (Schlyter 2003), is the system of advertising. The wide availability and knowledge of alternatives is a factor in facilitating people’s management of their housing asset.

In terms of actually accessing and retaining a unit, at Hillview, a prospective tenant simply puts down the upcoming month’s rental before the end of the month to secure tenure for the next. If the tenant cannot pay, she/he must move out. Likewise, if the tenant wishes to leave, she can do so at any point, but preferably informing the mastanda with some days to spare before the start of the next month, so she can find another tenant. Although there are no written agreements, it seemed that occupants generally seemed to abide by this etiquette and that it made sense to people (including the primary author while engaging in participant observation).

Nevertheless, there were cases where occupants felt unfairly ‘cheated’ by management in the process of securing a room or space. An example of this is where one participant paid a mastanda in advance for to occupy a full room he had viewed. When this prospective occupant arrived with all his belongings on the back of a hired bakkie, he found that the previous occupants had refused to move out as they were unable to find the money to do so, and yet the mastanda claimed to have already spent the money the prospective occupant had paid. The occupant was thus forced to share the room for the month, and move to a room nearby at the start of the next.

In literature on rooms and spaces, terms of tenure vary from monthly to much longer. In cases in Ghana (Tipple and Willis 1992) and Tanzania (Cadstedt 2010), leases are paid up front, months or over a year in advance, likely reinforced by rent control and family obligations (Rakodi 1995). In Thika, on the other hand, almost all tenancies are month-to-month (Andreasen 1989).Footnote 6 None of the literature on inner-city Johannesburg or the findings of the current study revealed rooms or spaces on the informal market that required more than a month’s rent in advance. In general, the type of tenure households consider appropriate to their needs varies widely (Payne 2002). In some cases, the stability of ownership is sought, in others, households choose flexibility so as to respond to livelihood opportunities if they arise elsewhere (ibid). In most cases, it is difficult to disentangle the relative significance of positive choice and negative constraints (Rakodi 1995), that is, whether people are choosing the tenure type that works for them best or are forced to simply take what they are able to get.

There are some cases where monetary rent is not paid. In the current study, a few occupants were able to secure a place on compassionate grounds or through relations (romantic or otherwise) with another occupant in the room or flat. When the lead author moved into Hillview, a young woman that had been staying in an abandoned car in the building’s parking garage asked the lead author’s mastanda if she could stay in the flat. The mastanda asked the other occupants if they would mind, and they agreed. The woman kept a small box of her possessions in the corner and slept there at night without being asked to contribute to rent (Fig. 4).

Fig. 4
figure 4

Some people decorated and set up very homely spaces, even though the floor area was limited. This was often based on how long they saw themselves wanting to or being allowed to stay in the particular room or space (source: author’s own photograph, August 2012)

Full size image

In Donkela, the barriers to entry into the building were quite high: if people did not know someone in the building who would let them stay with them for a while, they would generally have to ‘buy’ a room from the household moving out, and rooms were in such demand that occupants did not need to advertise. After an initial lump-sum payment, usually of about R600 to R900, it was understood that the occupant ‘owned’ the room. The R100 ongoing payment that each unit paid to the building committee was mostly seen as a levy specifically for the cleaning and security services.

In Donkela, there was a relatively low turnover, and many research participants had themselves stayed there for years, some since they had first arrived in Johannesburg. Even if they failed to generate an income, the minimal or zero cost of accommodation in Donkela meant they could continue occupying their unit. Furthermore, even when rooms became available, knowledge of the vacancy was minimal amongst the general public.

In the case of Abney, the most formal of the three buildings, people seeking a room were directed to the client relations department at management’s head office and were put on a waiting list to be notified about any vacancies. As rooms were mostly fully occupied, accommodation was not advertised. Before renting at Abney, potential tenants were required to put down 2-month deposit and sign a written lease indicating a minimum term of 6 months. They had to produce salary slips and bank statements proving that they were earning a steady amount of at least three times the rent, and the client relations department checked a database used by most large letting companies for records of conduct and payment. Abney had additional requirements, including proving one’s South African citizenship with an official ID book. These checks thus excluded as tenants those people that may have been able to pay using their informal income generating activities but did not have the documents to prove it. Even so, some participants managed to get around these restrictions by having a compliant family member sign the lease. For example respondent I14’s sister signed the lease although I14 paid most of it, and in return, I14 looked after that sister’s daughter: they partitioned the room using cupboards to make two spaces. Thus, most Abney occupants also remained in stable occupation, but for different reasons to those at Donkela, seemingly because they were happy with the accommodation and generally had more stable (often formal) incomes than those at Hillview or Donkela.

In this research, rooms and spaces involve a combination of ownership and non-ownership agreements that stretched across the informal/formal spectrum. Donkela offered occupants ‘ownership’ of their rooms, even though the entire property was formally owned by another company and the occupation of the building was unauthorised. Hillview primarily offered month-by-month rental, although main leaseholders were sometimes bound by formal written leases involving deposits and minimum terms, and some occupants were also able to secure accommodation without financial costs. In Abney all rooms required deposits, written agreements and suitability assessments. Most Donkela occupants would not meet the financial requirements of Hillview, and many Hillview occupants would not meet the additional requirements at Abney. Occupants of Abney would not want to stay in Donkela, because of the difficult conditions there. Occupants could be said to be making the best choices regarding limiting their vulnerability, given their constraints. As implied by Rakodi (1995), it is difficult to assess to what degree this is asset management or, particularly in the case of Donkela, simply opportunism.

Eviction

In other studies, it was found that the lack of formal lease agreement may not be critical in itself as the perceived and practical risk of eviction may be negligible even without a lease agreement. Thus, participatory assessments are crucial to understand vulnerability to eviction, rather than simply researching whether occupants have a title deed (Payne 2002). Schlyter (2003:55) reports that the verbal agreement utilised in multi-habitation ‘implied mutual respect for unspoken rules and social peace’, and renting in this way still had the potential to offer similar securities to ownership (Carey 2009). However, Grant (2007) describes how, in some cases, the absence of regulation may increase vulnerability for poor households: owners can arbitrarily raise rents, charge females extra or demand that lodgers must leave at short notice. Even so, many studies found that, despite the potential insecurity, occupants were only evicted in the case of non-payment or significantly breaking the landlord’s rules (Carey 2009). Poulsen (2010) found in Johannesburg that, if the rent was late, the occupant’s belongings were removed and he or she was locked out until it was paid. If this took too long, the room was rented to someone else (ibid).Footnote 7 Lemanski (2009) found that the lack of regulated agreements contributed to backyard dwellers being vulnerable to the whims of the landlords, yet they perceived themselves at less risk of eviction than residents of informal settlements. This was in part because they were ‘merged’ with households who held formal tenure rights (Lemanski 2009:473).

These aspects were similar to the findings in the current research. At Hillview, none of the occupants felt vulnerable to eviction by the mastanda or owner, as long as they paid their monthly rent and abided by the rules agreed to with the mastanda. Hillview seemed to have the most direct causality between non-payment and eviction: generally if one did not pay by the 7th day of the month, it was considered fair to evict; however, omastanda had different levels of leniency, and a number were reported to be understanding toward occupants who were unable to pay due to hardship. Interestingly, the lack of written agreements and lack of application of the Prevention of Illegal Evictions and Unlawful Occupation of Land Act (1998) and linked rulingsFootnote 8 seemed to result in additional flexibility for both the occupants (lessees) and the mastanda or owner. Nevertheless, amongst the three buildings studied there were some practices which people felt were unfair. Donkela was perceived to have some bias resulting from the committee apparently choosing ‘easy targets’ to set an example around evictions, and the strength of the committee occasionally being called into question with competing groups.

In cases of eviction at Donkela and Hillview, it was the rent collectors (the mastanda in the case of Hillview or the residents committee in the case of Donkela) who also performed the evictions: they had the freedom to do so because, generally, neither the omastanda nor the rest of the occupants, and often not the owners, knew or utilised the law directly, as they seemed to acknowledge that their living arrangements contravened the legal system in some way (regarding overcrowding, for instance). Nonetheless, many occupants did have some idea that their rights protected them from unfair eviction, which provided some kind of cushion against vulnerability as they could argue these with omastanda. At Abney, however, evictions involved additional processes and parties: management spoke of a number of cases where occupants organised themselves and sought legal representation, and management itself operated within the law. These procedures proved time-consuming and expensive (see below), and management at Abney indicated that by extension, the company had to over time increase the rent for the other occupants in order to cover these costs.

Relative Importance to Occupants of Different Aspects of the Accommodation

Services

Our third theme discusses the relative importance to occupants of different aspects of the accommodation. Access to functional services was possibly the most valued physical aspect of rooms and spaces—more so than the nature of the building, internal partitions, or what amenities people shared or not. The level of services is also used by Médecins Sans Frontières as a good indicator of living conditions in inner-city Johannesburg buildings (IRIN 2011). The services prioritised by respondents accord with those in other studies: access to toilets, running water, drainage and electricity. In each of the buildings, occupants shared toilets, water (for washing, cooking and cleaning) and drainage (for pouring dirty water) but used their own electricity supply, particularly for TVs or radios and charging their cell phones. In buildings without a refuse collection service, there tends to be a build-up in communal spaces (or public spaces, such as the street) and there was no evidence of occupants making a plan for private collection such as shown by Huchzermeyer (2011) in Nairobi.

Reliability in supply of services is key: when water or electricity supply cut out, it made people more vulnerable in terms of security (when the building and/or street was completely dark at night, for example), it caused increased and unplanned for costs of living (as people had to buy water, or buy more expensive cooked food when the electricity cut), and difficulties in heating and lighting with implications on ability to work or study at night. Stability in costs of services from month to month is also critical to preventing an increase in vulnerability: participants had a very limited discretionary income with which to buffer shocks in expenses.

The case of Donkela echoed the inner-city buildings discussed by Few et al. (2004); IRIN (2011) and Silverman and Zack (2007) where occupants experienced severe difficulties when basic services were not reliably in place. The availability and quality of these services are also closely linked with occupants’ perceptions of maintaining the health and hygiene of a building, discussed below.

Health and Hygiene

In the literature reviewed, hygiene and health differed greatly between case studies and seemed related less to the physical typology than to the type and condition of the property and, particularly, the effectiveness of management. None of the studies reviewed had extensive empirical evidence as to how occupants’ health had been compromised: Grant (1996:251), for instance, suggests that difficulties associated with sharing would all be ‘detrimental to health conditions’, but does not expand on this statement.

Participants in this research conceptualised ‘health’ in various ways, but most related it to cleanliness and hygiene in the building. This was much easier to maintain in the separate room typology of Hillview and Abney than in Donkela, mostly because of the nature of the more flimsy internal partitions, and the lower level and reliability of services in the latter. Occupants of Donkela with rooms in the corridor behind the security gate considered it one of the safest, cleanest areas in Donkela, as they self-managed the space, had access to who came and went there, and there was a greater feeling of neighbourliness. An unhygienic environment impacted severely on occupants’ well-being, and in Donkela, many participants felt vulnerable to illness and even death due to the adverse conditions, echoing Silverman and Zack’s (2007) caution that high density in badly managed environments such as ‘bad buildings’ can facilitate the rapid spread of disease and fire.

Abney had communal toilets, bathroom and kitchen for each floor although they were not physically restricted for use by anyone in the building. Some Abney occupants complained about the unhygienic state of the communal ablutions, while no one at Hillview did. Firm management of the individual units (and flats in Hillview) was shown to be essential for keeping a hygienic environment. Virtually all the flats were clean (well-swept floors, disinfected bathrooms and tidy kitchens), even if they were not well maintained. There was often a clear internal system in place to ensure this:

The rules were we should clean in time, in a week we could divide the days…we couldn’t give men to clean…in fact we were the four women that did it. We could clean 2, 2 days each in the whole flat. (Interviewee I9, 12 August 12)

In flats where there were both men and women occupants, the women generally cleaned, while men bought the cleaning products. In flats where it was only men or only women, it was a bit more versatile, and generally everyone helped to clean, but usually the decision as to the details of the system would be taken by the mastanda.

The lead author’s experiences in all three buildings were that children and many adults seemed to treat the communal spaces such as external passages, courtyards, etc. as not being their concern. The compound effect of this was a direct negative impact on everyone’s vulnerability through a build-up of rubbish. In Donkela, the dirt, human waste and rubbish was often one of the first things people mentioned as a negative aspect of life there, particularly when discussing health. It was partly for this reason that the Donkela residents’ committee began collecting a monthly levy to employ cleaners and security; this seems unavoidable additional expense where there are communal spaces shared by many people where it is possible to mess and remain anonymous. However, as is evidenced by the cleanliness of the shared self-contained flats in Hillview, if sharing occurs on a smaller scale between only those people with keys to a particular space, it is possible to pinpoint who caused the mess or did not clean and it is not necessary to employ a cleaner (Fig. 5).

Fig. 5
figure 5

Outside passageway leading to side of Donkela warehouse where informal recyclers sorted the collected materials. The French drain used for urinating and throwing dirty water (which then flowed into a stormwater drain) was also located here (source: author’s own photograph, September 2012)

Full size image

There was a periodic build-up of rubbish outside the building. One had to walk over this area with stagnant water and excrement in the dark if one did not want to urinate in a bucket in the room at night. Cleaners’ job descriptions were not clear about responsibility for external areas.

Safety and Security

A high level of security was also very important to occupants. With respect to Abney, even though occupants saw the immediate neighbourhood of Hillbrow as a dangerous place, one of the key reasons for choosing to stay in Abney was the tight security. It was important that the building was completely secured from people outside it: although Abney was large, participants did not feel threatened personally by other occupants and were only concerned about the theft of small items such as cell phones when they left the door unlocked while out of the room. In Donkela, despite people also not trusting other occupants of the building, many were willing to pay for and valued having a security guard at the gate to the exterior, particularly at night.

While Hillview occupants regretted not having tighter security for the building as a whole and security guards were not employed, participants felt safe within their flats, behind lockable doors or security gates with a shared key. Once again, the nature of the internal partitions demarcating the subunits of rooms was also important to people in feeling secure: people in spaces indicated they envied people who lived in their own rooms. However, some still sacrificed this opportunity, if it meant a big cost saving.

Once again, many people were prepared to subject themselves to what they perceived as greater vulnerability than at ‘home’, in order to have money to send home, to save or spend on alternatives, or simply to be able to support themselves.

Space and Privacy

Rooms and spaces, in many cases, would not exist if the owners or main leaseholders of the larger unit did not decide to sublet a portion. The practice appears widespread in other contexts: in a study in a neighbourhood of Bogota, 57 % of owners with more than one room rented out space, and in a neighbourhood in Lusaka, one third did so (McCallum and Benjamin 1985). In inner-city Johannesburg, most owners who chose to reside in their flats decided to sublet; 73 % to strangers they did not know before (Carey 2009; Gordon and Nell 2006). Studies confirm that the private rental sector is essential for the livelihoods of many small-scale landlords (Cadstedt 2010; Carey 2009). Many households rely on multiple modes of livelihood (Grant 1996) and rental income can be the most consistent of all streams (Schlyter 2003). Grant (2007) identifies how, when households faced increasing vulnerability such as the shocks of retrenchments and high inflation in Zimbabwe, even lodgers sublet their limited space to ‘provide some insulation’ against poverty (Grant 2007:83).

In inner-city Johannesburg, Few et al. (2004) notes that the number of people per rooms or space ranged from 3 to 12. Indeed, the oft-mentioned overcrowding threshold of two persons per room (UN-Habitat 2003) seems to be readily overlooked by poor urban dwellers (Follain et al. 1982). Schlyter (2003) notes a study of rooms and spaces in Zimbabwe by Withers (1992) where only a third complained about space, and half even claimed to be satisfied with their accommodation.

Even so, although occupants prioritised other aspects over personal space, a number of studies showed it was a luxury they wished for. Respondents in Schlyter (2003) desired separate bedrooms for parents, boys and girls. Furthermore, living in close proximity to each other and sharing sometimes limited facilities seemed to exacerbate any social dissonance. Grant (1996) describes how sharing to the extent found in some lodging houses increases everyday stress and affects the ability of the household to remain unified. She found that, for adults, the lack of privacy strains sexual relationships, and, for children, can create adverse conditions for studying for school (Grant 1996).

Contrary to official regulations, many occupants of rooms and spaces seemed to sacrifice the physical asset of living space in favour of another asset, particularly an additional income stream, for stability purposes, or in other cases to invest in growing their physical asset of housing over the long term (Dewar 1997).Footnote 9 Participants generally seemed to take on additional people to share spaces when there was even a small amount of floor space to spare, particularly with the system of exchange and management at Hillview Mansions. Many occupants also build up social capital with each other. Importantly, Schlyter (2003:64) highlights that the feeling of congestion has more to do with the social relations and the design of the space than with the number of persons per space unit. Again, it would seem that effective management is key.

Participants used innovative means to maximise storage space where floor area was limited. Walls were used to hang various items from washing basins to clothes, beds raised high on bricks with crates to store things beneath, and stackable, pest-proof paint buckets used to store food and consumables. The bed provided the primary place where occupants and visitors sat and watched TV, ate and socialised, and where small babies slept and played. After being made up for the day, the bed was essentially the ‘living area’, replacing the need for additional seating or furniture aside from storage space.

Some participants indicated that they utilised semi-public spaces within the buildings and public spaces without, to socialise, relax and literally get some space. At Hillview, the courtyard and parking garages were heavily used by children, and the rooftop by teenagers and adults. People gathered for instance in the shebeen (informal tavern) or street at Donkela, on the steps and corridors atHillview, or parks nearby.

It was clear that, in general, occupants prioritised finances over space, confirming similar findings by Follain et al. (1982); Lim et al. (1984) and UN-Habitat (2003). Sometimes the rooms and spaces occupied were considerably smaller than the minimum floor areas stipulated in the city of Johannesburg bylaws (2004b) and would be considered as overcrowding and thus illegal. However, as found in the literature (for example Schlyter 2003), people considered smaller spaces feasible living spaces.

Privacy was conceptualised differently by different participants, and occupants employed various tactics to manage, given the situation of living in such close quarters with others. Some made close friends, some purposefully did not. Either way, personal space was generally respected, even where the division between that and communal space was only a curtain. Thus, while occupants of rooms were envied by those with smaller subportions of space, these spaces still allowed occupants a good degree of visual and physical (if not aural) privacy. For some occupants, particularly men, the issue of privacy was not a particular concern.

There were subtle arrangements governing demarcation of personal space. In Hillview, it seems that whether one can put up curtain divisions can sometimes depend on whether one has paid for the space or not. Those who had been granted the place to sleep and store some belongings as a favour rather than for payment often did not put up curtains: they did not seem to have the same right as those who had paid and apportioned that area as ‘theirs’.

Some spatial arrangements happened more organically, according to the preferences of the occupants at the time (rather than the mastanda/owner) and were based on the household makeup. For instance, respondent I11 rented a room in Hillview Mansions with three friends. Originally, they had a more open arrangement, with only curtains separating the ‘bedroom’ spaces from the communal space. However, when I11 got a more permanent girlfriend, he negotiated a rearrangement of the room using wall units and a different curtain layout to create more privacy.

‘Anti-Social Behaviour’

Despite references to problematic behaviour generally linked to drinking in or next to the accommodation (Bank 2007; Schlyter 2003) aggregate reviews (UN-Habitat 2003; Rakodi 2002) suggest that generally positive, mutually beneficial relationships exist between occupants, including live-in landlords like the omastande of this study. However, many of the participants in Donkela highlighted the main problem of staying there as being the culture of heavy drinking and partying, which they said led to increased noise levels, theft, violence and abuse. One interviewee said, ‘living here is stopping being able to think’ (interviewee I1 14 September 12). There was a unanimous sense that this was worst between Friday and Sunday and impacted greatly on residents’ safety and security, with some fearing to go out of their room. Some participants blamed the setup of the building and its management for the development of this culture:

Staying here encourages people to drink…in a way, we forget about problems. If I stayed somewhere else I wouldn’t enjoy drinking—a place makes you [drink]…You can see this place—the stresses—most people don’t even have money for food but for booze there is [money]. I think it’s some sort of depression or something like that. (Interviewee I4, 23 September 12)

Occupants related this drinking and partying culture to further rule-breaking and violence: ‘people mustn’t urinate by the steps, or all over, but…most especially on weekends everyone just does it’ (interviewee I4, 23 September 12). The physical characteristics of the warehouse (rooms open to the roof, made with relatively flimsy material) also meant people were affected by this ‘anti-social behaviour’, even when in their own room.

Schlyter (2003) and Bank (2007) describe very difficult experiences related to the operating of ‘shebeens’ on the same property as the lodging house and backyard shacks.Footnote 10 Schlyter (2003:33) says, ‘it was a hard life. No one in the house could sleep until the morning, when the drinking stopped…most of the lodgers were single people who themselves participated in the drinking’. Bank (2007:221) found interpersonal violence and ‘anti-social behaviour’, which was also associated with the ‘excessive alcohol use’ prevalent in the yards studied.

In Hillview Mansions, certain units were known as the ‘place to go’ for a party and only people who enjoyed that kind of lifestyle rented a room/space there. However, because management occurred on a unit level as well as a building level with formal partitioning between the old flats, the other units were not as affected, and there was not a ‘culture of deviance’ across the whole building, as in Donkela. In Abney, due to the more formal and separate physical characteristics and tighter management combined with more people with full-time jobs, these occupants also did not experience the social issues found at Donkela.

Discussion

This research revealed the incredible resilience people seeking a livelihood in Johannesburg’s inner city in putting up with sometimes difficult living circumstances. Participants in this study were often reliant on transient livelihood opportunities, and occupants of rooms and spaces have managed to secure a form of accommodation that supports their livelihood strategies as best as possible. The livelihood contributions of the accommodation include, amongst others, the good location, significantly lower accommodation costs as a result of subdivision, the flexibility, mobility, diversity, limited threat of ‘unfair’ eviction, and in the Hillview and Abney buildings, the reliable services, and the healthy and safe direct living environment that was reported by participants.

Nevertheless, the shocks and stresses occupants had to endure in some situations severely undermined livelihoods. This was particularly evident in Donkela, and to a small degree, in some cases in Hillview, particularly for occupants of spaces as opposed to rooms. These stresses included pressure on relationships as a result of the high degree of sharing and limited space, and some threat of eviction in certain cases. Donkela occupants faced the greatest additional vulnerabilities, as a result of the physical characteristics of that specific form of rooms and the warehouse, and the actions of their fellow occupants. Occupants could become ‘trapped’ in the ‘culture’ of anti-social behaviour and were subject to a great deal of shocks and stresses resulting from the lack of services, severe violence at times and general lack of safety particularly for women, serious health threats and hygiene problems, and more. There is definitely scope for intervention in buildings such as these.

Thus, overall, this research found that, while occupants in Donkela were faced by the most tenuous livelihoods, they lived in the building that, out of the three studied, resulted in the most additional shocks and stresses. Hillview occupants had been more successful in securing livelihood opportunities, and many aspects of this ‘model’ of rooms and spaces were of great support to these livelihood strategies, though with some negative implications. Abney occupants had the most secure livelihoods, and staying in that building also resulted in the least additional shocks and stresses.

It is possible to highlight some more specific lessons for practice that can be taken forward from the three different forms of rooms and spaces. Donkela provided by far the lowest cost accommodation and still had a semi-functioning management system. However, it had the potential to more severely undermine occupants’ livelihoods than support them. However, depending on an occupant’s location within the building and particularly ones current health, abilities age and gender, some occupants were much less vulnerable and were able to benefit from the low cost and good location of the building. Nevertheless, for the most vulnerable in the building, particularly women and children living on the lower levels, if there were any possibility of alternative accommodation being available, their needs should be prioritised.

Rooms and spaces in Hillview Mansions were more supportive of the occupants’ livelihoods in general. The management system and availability and general structure of the internal services on a per-flat basis allowed some easing of the social tensions and physical pressure that the high densities and mobility had the potential to cause. Nevertheless, occupants of spaces, particularly those sharing spaces, faced far greater vulnerability than those in rooms. No matter how small the room was, the physical characteristic of having a sealed, closable area to sleep and store one’s possessions made a very big difference, and participants unanimously valued living in a room over a space. Again, for women and children, and particularly women with young children who were mobile but not yet disciplined enough not to stray into other spaces, who cried in the night, and other issues, this was more important. Single men were more inclined to stay in a space (rather than a room) and spend the money saved on other things. This sometimes created tensions between male and female partners. To improve conditions, design interventions could be investigated here, such as using floor to ceiling dry walling to split the larger rooms. Indeed, this has already been done on an informal basis by some omastanda and on a formal basis by large property companies (where drywall rooms and flats rent for roughly the same as those partitioned using brick and mortar). In this way, ‘clustered units’ could be formed, with smaller rooms clustered around shared washing and toilet facilities. The ‘self management’ system found in Hillview Mansions could definitely be piloted in this setup.

The mastanda system of management was essential to maintaining many of the livelihood supports that the Hillview Mansions rooms and spaces provided, and lessening the shocks and stresses. One of the key virtues of this system is that it operates without explicit government or external intervention. In the current policy gap around very low cost, compliant rental accommodation, it seems clear that this largely functional system should be left to do so. It would be difficult to ‘artificially’ set up this system in other buildings which currently have a formal system, but elements of it could certainly be borrowed, for instance, allowing the main lessee, the mastanda, the freedom to sublet and thus drop the cost of accommodation to fulfil demand and introduce greater flexibility through the formal system, etc. Here, incentives for self-management within the units could develop like in the Hillview model, where omastanda keep a well-functioning and fully occupied set of rooms or spaces through being able to discount their own accommodation costs. Peoples’ practices at Hillview and similar buildings were closely linked to the physical characteristics of rooms and spaces found there, and thus, these aspects could be borrowed as well. This scale of management could also be kept to similar ratios through physical design of the accommodation, i.e., one live-in mastanda per around eight to 16 occupants (which corresponds with the sharing of common facilities).

Abney presented a situation that supported occupants’ livelihoods well. For those that were able to afford to stay, it worked relatively well. However, management limited which people were able to stay there based on proof of income and other means tests, and the lower level of subdivision allowed meant the rooms and spaces were less affordable than the other two buildings studied. Moreover, borrowing from the physical configuration and management structures found in buildings such as Hillview when rolling out formal rooms for rent may well provide greater livelihoods support for occupants, lowering management costs, increasing affordability and allowing for greater flexibility.

One of the key findings of this research is that occupants across all three buildings seem to prefer cutting accommodation costs while maintaining certain key aspects to the accommodation, rather than retaining more living space. These key aspects include safety and security, hygiene and particularly access to basic services, even where shared by many people within the larger unit. This should be kept in mind in new developments, if costs would be saved by reducing the space available.

However, rooms and spaces are not always viewed in a favourable light by government, particularly those typologies on the informal market. There is an extreme shortage of cheap rental accommodation in the city which meets local authority standards, and the subsequent range of illegal or unsanctioned practices cause a crisis for the city management. Even though legal precedent based on the South African ConstitutionFootnote 11 protects occupants from eviction without the provision of alternative accommodation should they be rendered homeless, efforts to eliminate ‘bad buildings’ are often lauded as ‘inner-city regeneration’ and ‘rejuvenation’ by government and in popular discourse at home (see eProp 2012) and abroad (see Curnow and Joy 2013), yet the implications for the evictees are often overlooked (COHRE 2005). Moreover, current legislation and policy poses some restrictions on rooms and spaces that would reduce ore remove some of the key livelihood supports found in the varieties of rooms and spaces studied, particularly those at Hillview Mansions. Regarding physical characteristics, City of Joburg public health bylaw (2004b) restricts the density of rooms and spaces to rising to the levels found in some of the flats in Hillview Mansions, as well as the use of spaces as sleeping areas not designated as such. The Problem Properties bylaw (City of Joburg 2014) enforces these elements, makes additional reference to the contravention of town planning schemes and building regulations, and allows for the building to be placed under administration should the owner(s) not comply following a notice.

More worrying is the implications for peoples’ practices of the Rental Housing Amendment Bill of 2014, which, amongst other aspects, demands all lease agreements to be formalised in writing according to uniform government-stipulated requirements. The same legal precedent that protects certain occupants from eviction has meant that some landlords are even stricter with their entrance requirements for new lessees, and limit mobility through the requirement of longer term leases, as the process required to evict occupants should they refuse to pay rent proves far too expensive. Lastly, the requirements for Community Residential Units and Social Housing to be self-contained units as per the National Housing Code (National Department of Human Settlements 2009) mean that the associated capital subsidy allocations are not available for developers wishing to provide rooms or spaces. In this regard, there is a general fear within government of creating housing that is in any way similar to the Apartheid-era hostels (which were a badly constructed and managed form of rooms and spaces).

Moreover, intervention by no means has to take place only within the realm of housing. Few et al. (2004:439) emphasise the need to address the complex nature of the housing crisis in developing country cities: ‘the housing problem is not something that can be solved by architects and planners alone’. The findings of the current study suggest that it is tenuous livelihood opportunities that are the root of participants’ vulnerability: if they were to secure stable livelihood opportunities, many of the livelihood supports that rooms and spaces provide would be unnecessary.

Conclusion

This paper has documented the widespread phenomenon of rooms and spaces as studied in three different buildings in Johannesburg’s inner city. A key aim was to describe and analyse these examples in depth in order to understand better this neglected object of study, particularly in light of the existing gap in housing policy response to the intense demand for cheap rental accommodation. We found that rooms and spaces can assist occupants to minimise expenditure on rent whilst maximising locational advantage and are also able to offer flexibility to allow for adaption to changing circumstances. These features suit the insecure and fluctuating livelihood strategies of many of the participants.

The three buildings illustrated some of the diversity of physical characteristics that can be captured under the term ‘rooms and spaces’, and demonstrated that some features act as an asset, while others in fact add to vulnerability. This highlights the need to understand specific physical and service conditions when reviewing the strengths and limits for the occupier of rooms and spaces.

The study found that access to reliable basic services (water, toilets, electricity and rubbish collection) was one of the highest priorities for most participants, and health and hygiene was closely associated and highly valued within the personal living space. Safety and security were prized, and both services and security were prioritised over additional floor space. People made innovative arrangements to cope with the minimal living space, using the bed almost 24 h a day, for living, eating, playing as well as sleeping.

Each of the buildings forming part of this study demonstrated a combination and layering of management and of tenure arrangements, with the two buildings with more informal characteristics (the subdivided warehouse Donkela and the informally converted units of Hillview Mansions) showcasing the greatest complexities. A mix of ownership, assumed ownership, formal leasing and formal management companies, multi-layered subletting, relatively formal levies in an entirely informal arrangement, constituted wide variety of typological and tenure offerings in many ways supported livelihood strategies and allowed access to people who would otherwise have nowhere to stay but ‘under a bridge’.

The key supportive features of rooms and spaces were found most commonly at Hillview Mansions, where the flexibility and mobility afforded occupants through the multi-layered rental and subletting arrangements, advertising and infrastructure around moving from place to place allowed occupants almost full flexibility on a month-to-month basis. However, even in the Abney building studied, where rooms had been developed on a formal basis, ‘under the radar’ subletting allowed occupants access to the accommodation who would not otherwise have qualified or afforded it, and greater flexibility for all.

In a context of no better housing alternatives, we suggest that a livelihoods perspective on rooms and spaces indicates that interventions into these existing conditions should be approached with caution, lest they worsen lives that are already precarious. It can be argued that some of the rooms and spaces typologies, even those unsanctioned by city authorities, are ‘working’ for their occupants, and thus Carey’s (2009: 2) caution to ‘do no harm’ needs to be taken seriously whilst acknowledging the city’s obligation to ensure occupants and wider society are not placed further at risk. From a policy perspective, Berner (2001:302) suggests that ‘what is needed, in brief, is a bridging of the gap between the legal and the illegal systems, starting with the recognition that illegal systems have in the past achieved far more than any official initiative’.

In this research, the building, Hillview Mansions, demonstrated a particular (informal) model of rooms and spaces that was well adapted to peoples’ unstable livelihoods, allowing occupants to adapt the accommodation to their circumstances without being displaced in the building. The research was able to uncover the complex combination of elements that allowed this flexibility and potential for mobility, including the wide range of differently priced and laid out units (subunits of the original flats) from which to choose, localised mastande management system in combination with formal management structures, multi-layered subletting, informal tenure relationships including less strict application of eviction law yet very limited practices considered unfair by occupants. This is a model that could offer potential learning for future housing strategy. On the other hand, the findings suggest that occupants of the Donkela building may do well to have an intervention: the negative impacts on their livelihoods outweigh the possible benefits. Allenby worked relatively well for those people that could afford the accommodation or qualified, but the relative lack of flexibility and possibility of mobility, as well as more centralised management structures would exclude most occupants of Donkela or Hillview Mansions.

While the current research focused on three specific buildings, by following different occupants as they moved to different places over the course of 2 years after the conclusion of interviews, and work carried out with occupants of additional buildings, the lead author was able to get a relatively thorough internal view of around 20 other buildings demonstrating rooms and spaces in Johannesburg’s inner city. The variety of rooms found in Donkela were common in other warehouses, while the informal subdivisions of flats at Hillview Mansions were very widespread in blocks of flats both formally less formally managed. The formal rooms at Allenby were common in some formally converted and managed buildings, particularly those run by social housing institutions. As mentioned previously, many of the findings in the current research correlate well with existing literature (e.g., Carey 2009; Few et al. 2004; Poulsen 2010). Indications are the findings from this study are likely to extend to other similar buildings, however, there is an urgent need for a study at a broader level of buildings utilised for accommodation in Johannesburg’s inner city to enable the broadening of the recommendations of this study. An inner-city housing action plan as proposed by the municipality’s inner city roadmap (City of Joburg 2013) would do well here. Furthermore, currently, we are seeing a continued divergence between legislation and case law on the one hand, and what arises organically and informally on the other, including what supports occupants the most given their realities. This makes regulation of accommodation even more difficult, and a greater divergence between what authorities turn a blind eye to, compared to formally managed and more tightly regulated accommodation. There would need to be a reexamination of legislation governing the management/leasing relationships (Rental Housing Amendment Bill 2014) and physical characteristics that are contravened in rooms and spaces found in buildings such as Donkela and Hillview Mansions, with due consideration to fire and safety concerns. The findings of this research show that rooms and spaces vary significantly in nature and have varied and important impacts on livelihoods. Ignoring, criminalising or dismissing significant components of existing practice is not strategic, and case-based or stratified, targeted responses are likely to be an essential part of a future housing strategy.

Dates of Primary Interviews Cited in the Text

  • I1, Donkela, 14 September 2012

  • I4, Donkela, 23 September 2012

  • I5, Donkela, 23 September 2012

  • I6, Donkela, 4 October 2012

  • I7, Donkela, 4 October 2012

  • I9, Hillview Mansions, 12 August 2012

  • I11, Hillview Mansions, 15 August 2012

  • I13, Hillview Mansions, 18 August 2012

  • I14, Allenby, 12 August 2012