A NATIONAL SURVEY OF GATED COMMUNITIES IN SOUTH AFRICA

Transcription

1 A NATIONAL SURVEY OF GATED COMMUNITIES IN SOUTH AFRICA Prepared by: Karina Landman CSIR Building and Construction Technology

2 A NATIONAL SURVEY OF GATED COMMUNITIES IN SOUTH AFRICA Document prepared by: Karina Landman CSIR Building and Construction Technology (Boutek) Programme for Sustainable Human Settlements Assisted by: Renee Koen and Martina Tswai (Statistical programming and generation of basic tables for analysis CSIR icomtek) Peter Schmitz (Production of GIS maps on crime and income distribution CSIR icomtek) Lynette Smit (Production of maps on the distribution of gated communities CSIR Boutek) Telephone: (012) Facsimile: (012) Date: 2003 Project Number: BP 565 Boutek Report Number: BOU/I 257 Cover photographs: Top: Enclosed neighbourhood in Johannesburg Bottom: Security village (townhouse complex) in Pretoria CSIR Building and Construction Technology P O Box 395 Pretoria 0001 South Africa

3 CONTENTS CHAPTER 1: INTRODUCTION 1.1 Background 1.2 Structure of the document 1.3 Crime patterns in South Africa 1.4 Income distribution 1.5 Institutional and spatial context of study 1.6 Research methodology Objectives Research strategy, design and methods Survey response Analysis of data CHAPTER 2: CONFIRMATION OF TYPOLOGY 2.1 Working typology 2.2 Confirmation of working typology 2.3 Clarification of terminology regarding the typology and terms used in the questionnaire and survey analysis CHAPTER 3: DISTRIBUTION OF GATED COMMUNITIES IN SOUTH AFRICA 3.1 Overall distribution in provinces 3.2 Distribution of different types in municipal areas 3.3 Distribution of different types within cities CHAPTER 4: POSSIBLE REASONS OR RELATIONSHIPS FOR THIS DISTRIBUTION 4.1 Relationship with high crime areas Crime and gated communities in provinces Fear of crime and gated communities Crime and gated communities in municipalities Crime and gated communities in the Cities of Johannesburg and Tshwane 4.2 Relationship with higher income areas Distribution of income and gated communities in provinces Distribution of income and gated communities in municipalities Distribution of income and gated communities in the Cities of Johannesburg and Tshwane 4.3 Relationships of different types of gated communities with the types of municipalities or location of the municipality Urban/non-urban municipalities Sizes of urban areas in municipalities Metro/non-metro municipalities Coastal/inland municipalities Municipalities with popular tourist destinations/natural features CHAPTER 5: GOVERNMENT RESPONSE TO GATED COMMUNITIES 5.1 Policies on types of gated communities 5.2 Relationships between policy development, applications and approvals for types of gated communities and types/location of municipalities 5.3 Policies on gated communities in the Cities of Johannesburg and Tshwane CSIR Building and Construction Technology 2

4 CHAPTER 6: CONCLUSION 6.1 The location and extent of gated communities in South Africa 6.2 Explaining the distribution 6.3 The governments response to gated communities 6.4 Key findings and policy implications 6.5 Directions for future study or research REFERENCES APPENDIXES APPENDIX A: QUESTIONNAIRE APPENDIX B: COVERING LETTER APPENDIX C: VARIABLES IDENTIFIED FOR ANALYSIS APPENDIX D: CROSS-BOUNDARY LOCAL MUNICIPALITIES IN SOUTH AFRICA APPENDIX E: CROSS TABULATIONS OF ENCLOSED NEIGHBOURHOODS WITH OTHER TYPE OF GATED COMMUNITIES IN SOUTH-AFRICA APPENDIX F: CROSS-TABULATION OF ENCLOSED NEIGHBOURHOODS AND SECURITY ESTATES WITH PROVINCES IN SOUTH AFRICA APPENDIX G: LOCATION OF SECURITY ESTATES IN THE CITY OF CAPE TOWN APPENDIX H: POPULATION DISTRIBUTION IN THE CITIES OF JOHANNESBURG AND TSHWANE PER IDP REGIONS APPENDIX I: CORRELATIONS BETWEEN CRIME AND NUMBERS OF MUNICIPALITIES WITH GATED COMMUNITIES IN PROVINCES IN SOUTH AFRICA APPENDIX J: CORRELATIONS BETWEEN OVERALL CRIME RATES IN MUNICIPALITIES AND ENCLOSED NEIGHBOURHOODS AND SECURITY ESTATES APPENDIX K: CORRELATIONS BETWEEN THE SIZE OF CITIES IN MUNICIPALITIES AND ENCLOSED NEIGHBOURHOODS AND SECURITY ESTATES CSIR Building and Construction Technology 3

5 LIST OF FIGURES Figure 1.1 Changing crime rates Figure 1.2 Recorded crime, by category, Figure 1.3 Recorded crime in provinces Figure 1.4 Feelings of safety per province Figure 1.5 Recorded crime in selected police areas, 2000 Figure 1.6 Feelings of safety per area of residence Figure 1.7 Survey response per province Figure 1.8 Creation of new polygons (union) from two different layers. Figure 1.9 Crime data spatially displayed according to magisterial boundaries Figure 2.1 Definition and hierarchy of different types of gated communities in South Africa Figure 2.2 Public and private approaches to enclosed neighbourhoods Figure 2.3 Percentages of different types of gated communities in South Africa Figure 2.4 Distinguishing between road closures and an enclosed neighbourhood Figure 3.1 Percentage of local authorities with enclosed neighbourhoods Figure 3.2 Percentage of local authorities with security estates Figure 3.3 Number of local authorities with enclosed neighbourhoods and security villages per province (including only those who responded) Figure 3.4 The proportion of enclosed neighbourhoods in different parts of the city Figure 3.5 The proportion of security estates in different parts of the city Figure 4.1 Existence of enclosed neighbourhoods in high property crime municipalities Figure 4.2 Existence of large security estates in high property crime municipalities Figure 4.3 Recorded crime in selected police areas, 2000 Figure 4.4 Percentage urban population per province Figure 4.5 Distribution of municipalities with more and less than 50% urban population in South Africa Figure 4.6 Distribution of percentage urban population in municipalities in South Africa Figure 4.7 Comparison between percentage coastal and inland municipalities Figure 5.1 Policies on road/neighbourhood enclosures Figure 5.2 Year that policies were developed Figure 5.3 Municipalities in the process of revising or developing policies on road/neighbourhood enclosures Figure 5.4 Percentage of respondents with applications for road/neighbourhood enclosures Figure 5.5 Percentage of approvals of respondents with applications for road/neighbourhood closures Figure 5.6 Relationship between the number of applications and number of approvals for road/neighbourhood enclosures LIST OF TABLES Table 1.1 Table 1.2 Table 1.3 Table 1.4 Table 2.1 Table 3.1 Table 3.2 Table 4.1 Table 4.2 Distribution of local municipalities across provinces Research methodology of national scan Responses per province (metropolitan and local municipalities) Percentage urban population in municipalities and responses Distribution of types and sub-types of gated communities across municipalities Number of municipalities with enclosed neighbourhoods and security estates per province Location of enclosed neighbourhoods and security estates within cities in municipalities Division of violent and property crime types Description of selected information in the cities of Johannesburg and Tshwane CSIR Building and Construction Technology 4

6 Table 4.3 Table 4.4 Table 4.5 Table 4.6 Table 4.7 Table 5.1 Distribution of enclosed neighbourhoods and security estates per categories of percentage urban population in municipalities in South Africa Metro/non-metro municipalities with enclosed neighbourhoods and security estates Metropolitan municipalities with enclosed neighbourhoods and security estates Coastal / inland municipalities with enclosed neighbourhoods and security estates Distribution of number of security estates in coastal / inland municipalities Number of applications received per municipality LIST OF MAPS Map 1 New municipal boundaries in South Africa Map 2 The distribution of enclosed neighbourhoods in South Africa Map 3 The distribution of enclosed neighbourhoods in Gauteng Map 4 The distribution of large security estates in South Africa Map 5 The distribution of large security estates in Gauteng Map 6 Sizes of urban areas in municipalities Map 7 Total crime rates per province for the period Map 8 Violent crime rates per province for the period Map 9 Property crime rates per province for the period Map 10 Overall crime rates per municipal area for the period Map 11 Violent crime rates per municipal area for the period Map 12 Property crime rates per municipal area for the period Map 13 Total crime rates per persons for the period in the City of Johannesburg Map 14 Violent crime rates per persons for the period in the City of Johannesburg Map 15 Property crime rates per persons for the period in the City of Johannesburg Map 16 Total crime rates per persons for the period in the City of Tshwane Map 17 Violent crime rates per persons for the period in the City of Tshwane Map 18 Property crime rates per persons for the period in the City of Tshwane Map 19 Average income per capita (2000) per province in South Africa Map 20 Total income per municipality in South Africa Map 21 Total income per IDP Region in the City of Johannesburg Map 22 Total income per IDP Region in the City of Tshwane Map 23 Municipalities with applications for neighbourhood enclosures CSIR Building and Construction Technology 5

7 CHAPTER 1: INTRODUCTION 1.1 Background Gated communities are a global phenomenon. They occur in many forms in a large number of countries, including the USA, Canada, Brazil, Argentina, UK, Spain, Portugal, Bulgaria, Russia, China, Malaysia, the Philippines, Australia. There are also a number of countries in Africa that have gated communities, including Nigeria, Kenya, Zimbabwe and South Africa. During the past few years, gated communities have rapidly increased in South Africa. This has been the case for a number of different types of gated community. Although the major growth of enclosed neighbourhoods has occurred in metropolitan areas, especially Gauteng, the occurrence has not been limited to these areas. A number of applications for neighbourhood enclosures have also been received in municipalities with medium and smaller towns. In addition, the development of security villages is also escalating as this type of living place is becoming increasingly popular among urban residents. In the context of high crime rates, many people consider enclosed neighbourhoods or security villages the only option for safe living in cities. However, not all people agree that this is the only or best solution to crime in the country. There is intense debate surrounding the issue. CSIR Building and Construction Technology (Boutek) has been conducting research on gated communities in South Africa over the past two years. Gated communities 1 include security townhouse complexes, larger security estates, enclosed neighbourhoods, etc. The aim of this ongoing research is to provide relevant and accurate information to assist decision-making on gated communities in the country. Boutek conducted the first survey on enclosed neighbourhoods towards the end of 1999 and the beginning of As part of this study, questionnaires were sent to a sample of 100 local authorities in the country. There was a 43% response rate. At that time almost half of the respondents received requests for neighbourhood enclosures and were therefore forced to deal with the issue. Forty percent of these requests were approved. Generally requests were denied either because the local authority did not support the idea of enclosed neighbourhoods or because they foresaw problems with urban management and maintenance. Another reason for not approving requests was because no formal policy was in place. Neighbourhood closures were generally approved for security reasons and if there were no objections to such an enclosure (Landman 2000). Since then, the number of gated communities has proliferated in South Africa, making it necessary to get a clearer and more updated picture of the situation in the country, based on a larger sample and including additional questions. The first phase of Boutek s current research on gated communities comprised a national survey, and the second phase focused on in-depth case study research. This report will be concerned only with the national survey (Phase 1). Its main focus will be to look at gated communities as a new (or emerging in a new form) urban type and the impact that this urban type can have on future urban development in developing countries and more 1 A gated community refers to a physical area that is fenced or walled off from its surroundings, prohibiting or controlling access to these areas by means of gates or booms. In many cases, the concept can refer to a residential area with restricted access, so that normal public spaces are privatised or their use is restricted. It does not, however, refer only to residential areas, but may also include controlled access villages for work (office blocks), or commercial and/or recreational purposes (many shopping complexes, malls, etc). CSIR Building and Construction Technology 1

8 specifically in South Africa. The study will also provide updated empirical data to assist decision-making and policy formulation to this regard. International findings and local comments (newspaper articles), have indicated that there is often a strong relationship between crime and income. This document will also investigate whether this is true in SA and whether there are any other possible reasons for the distribution of gated communities in the country or within municipalities. 1.2 Structure of the document The main body of this document is divided into four chapters. Chapter 2 deals with the typology of gated communities in South Africa and provides a descriptive foundation for the next chapter by explaining the two major types involved in the survey. Chapter 3 discusses the distribution of gated communities in South Africa and distinguishes between provinces, municipalities and two metropolitan areas (the Cities of Johannesburg and Tshwane). The next chapter (4) takes this further, and explores possible relationships between a number of variables to determine motives behind this distribution or important influencing factors. The two main variables of concern are crime and income on the one hand, and whether these can be linked to the occurrence of gated communities in specific geographic areas, for example provinces, municipalities and even within municipalities (the cities of Johannesburg and Tshwane) on the other. Chapter 5 discusses the government s response to the demand for and growth of gated communities, and specifically to enclosed neighbourhoods and large security estates in the country. It shows where there are policies in place to deal with certain types of gated community, and investigates whether there is a link between policies in place and the demand for neighbourhood closures. Chapter 6 concludes the discussion by highlighting the key findings of the study and policy implications for government. 1.3 Crime patterns in South Africa Since the transition to democracy in 1994, crime has been one of the priorities that government has had to address and has often been identified as the top priority by residents in numerous surveys, community-participation programmes and workshops. Crime is also generally considered to be the main reason for the proliferation of gated communities in the country. A study of the changing crime rates in South Africa from 1994 to 2001 shows a decrease from 1994 to 1997, and then an increase from 1998 to 2000, surpassing the original 1994 rates (see Figure 1.1). There is a slight decrease from 2000 to One will, however, have to monitor these trends in the coming years to see whether this downward move is a true sign of stabilisation and an indication of a downward trend, or only a once-off variation. CSIR Building and Construction Technology 2

9 Number per people Changing crime rates, Jan-Sep Figure 1.1: Changing crime rates (recorded crimes) (Shönteich 2002b:1) The majority of recorded crimes were property crimes approximately 55% of the total. Therefore, while the total number of recorded crimes has increased since 1994, there has been little change in the proportions of the different categories (see Figure 1.2). If one considers only violent crime for example, this comprised 30.1% of all recorded crimes in 1994 and 32.5% of all recorded crimes in All other crime types experienced a slight decline as a proportion of the overall number of recorded crimes (Schönteich 2002a:2). Recorded crime, by category, Jan-Sept Figure 1.2: Recorded crime (Schönteich 2002b:1) Number Violent crime Property crime Arson & property damage Commercial crime Other The distribution of crime per province in 2001 more or less reflects the same proportions (according to categories) as for the overall distribution in the country from 1994 to What is however, interesting is that while the proportion of violent crime in Gauteng and the CSIR Building and Construction Technology 3

10 Western Cape is almost the same in 2001, the proportion of property crime is much higher in the Western Cape. Crime rate, Jan. - Sept N Province E Cape KZN North West Mpumalanga Free State RSA N Cape Gauteng W Cape Figure 1.3: Recorded crime in provinces (Shönteich 2002b:2) Recorded crimes per population Violent Property Damage to prop. Commercial Other The most recent national survey done on the fear of crime by the HSRC in 1999 (Figure 1.4) revealed that the highest percentages of people feeling unsafe were in the Eastern Cape (57%) and Gauteng (51%). Gauteng also had the lowest proportion of people (34%) feeling safe. One explanation is constant focus on violent crime incidents reported by the media. During 1999, Gauteng had the highest recorded per capita levels of attempted murder, robbery with aggravating circumstances and motor vehicle theft, and the third-highest per capita rates of rape, serious assault and murder (Louw 2001:2). 120 Figure 1.4 Feelings of safety per province Source: HSRC (cited in Louw 2001:3) 100 Percentages Neither Unsafe/ very unsafe Safe/ very safe Eastern Cape Free State Gauteng KwaZulu Natal Mpumalanga Northern Cape Limpopo North West Western Cape CSIR Building and Construction Technology 4

11 Above-average (the national average was 44%) feelings of safety were expressed by respondents in the Western Cape, Free State and the Northern Province (now Limpopo). In 1999, the Northern Province had the lowest per capita levels of 14 out 17 most serious and prevalent crimes reported. From the responses in the Western Cape, it however appears that perceptions do influence people s feelings of safety. The Western Cape is perceived to be a safe province (second-highest number of respondents feeling safe), but recorded crime figures show a different picture. During 1999 the province had the highest per capita rate of recorded murder, burglary and common assault, and the second-highest rate of attempted murder, rape, serious assault, theft of motor vehicles and common robbery (Louw 2001:3). Perceptions therefore also play a very important role in people s feelings of safety and their fear of crime. Crime patterns also differ for the five largest cities or urban conglomerations in South Africa. Recorded crimes in selected police areas in 2000 indicated that the highest rates occurred in Johannesburg, followed by Pretoria and then Cape Town (Figure 1.5)., leading to the heightened fear of crime in Gauteng, since Johannesburg and Pretoria, the largest cities, are both situated in that province. Recorded crimes in selected police areas, 2000 Figure 1.5: Recorded crime in selected police areas (Schönteich 2002b:3) Durban 6115 Port Elizabeth West Metrople (CT) Pretoria Johannesburg No. per population The high levels of violence in the large cities and metropolitan areas also explains the heightened fear of crime in these areas. According to the HSRC survey, over half of the respondents (53%) in rural areas felt safe, compared to urban residents, where just over 40% (43%) and only one-third in metropolitan areas (35%) felt safe (Figure 4.6). One explanation for the higher fear of crime in metropolitan areas is awareness of the actual crime rate. During 1999, 55 murders per population were recorded in South Africa as a whole. With the exception of Pretoria, the rate was higher in metropolitan areas: 117 in Johannesburg, 82 in Durban, 68 in Cape Town, but 37 in Pretoria (Louw 2001:1). Louw (2001) also points out that it is a worldwide phenomenon that certain types of crime are more prevalent in urban areas because of the presence of certain predisposing factors. For example, population density is thought to be associated with crime in that greater concentrations of people lead to increased conflict, greater levels of stress, and competition for limited resources. CSIR Building and Construction Technology 5

12 120 Figure 1.6: Feelings of safety per area of residence Source: HSRC (cited in Louw 2001:2) 100 Percentages Metro Urban Rural This discussion shows that the distribution of crime rates and different types of crime varies between different geographic areas, including provinces and cities. It also indicates that the fear of crime also varies and cannot always be linked to actual crime rates. The fear of crime is very dependent on the perception of crime, and often one violent reported crime, sensationalised by the media, can contribute to an escalation in the fear of crime. Generally, people in larger cities and metropolitan areas in South Africa are more afraid of being a victim of crime than those living in rural areas. The question is whether this stimulates increased and more extreme responses to crime in larger urban areas? 1.4 Income distribution in South Africa South Africa is regarded as an upper middle-income country (World Bank classification) in terms of its GNP per capita (1999), both according to the Atlas and the Purchasing Power Parity methods (World Bank World Development Report 2000/2001: htttp:// South Africa has one of the highest income-disparity coefficients in the world (currently fourth highest in the world). For a long time South Africa s Gini coefficient 2 (currently 0.58) was the highest in the world. One can also express income inequality by looking at the proportion of total national income of groups of households, arranged in order of income level. In South Africa the poorest 10% of the population receives only 1.4% of the total income, while the richest 10% receives 47.3%. Therefore, although it is generally regarded as a middle-income country, due to the large number of poor people it is in many cases considered a developing or transitional country. Lawrence (1990) calls these countries advanced third-world countries. 3 2 The Gini coefficient is one of the indicators for income inequality and it ranges from 0 (absolute equality) to 1 (absolute inequality). 3 Lawrence highlights a number of characteristics of advanced third-world countries. These include the size of the country, the presence of a colonial legacy, a relatively sophisticated economies, grossly unequal distribution of wealth, urbanisation, inequality of well-being or social inequality, and political uniqueness. CSIR Building and Construction Technology 6

13 The inequality in the distribution of wealth is generally considered a major contributing factor to the high level of crime and to reactive responses (including gated communities) to crime in the country seen since the transition to democracy in Institutional and spatial context of study As part of the drive for greater integration, improved efficiency and developmental local government after 1994, it was decided to re-demarcate all local and district authorities in the country and reduce the number of local authorities from around 820 to just below 250. In December 1999 the Demarcation Board of South Africa published the newly determined metropolitan, local and district municipalities and their borders, in accordance with the Municipal Demarcation Act (1998). A total of 284 municipalities has been demarcated: six metropolitan municipalities, 47 district municipalities, 231 local municipalities and a number of district management areas (mainly conservation areas) (see Map 1). The metropolitan municipalities are as follows: City of Johannesburg (in Gauteng) City of Tshwane (in Gauteng, including Pretoria) Ekurhuleni (in Gauteng, including East Rand) e Thekwini (in KwaZulu-Natal, including Durban) Nelson Mandela (in the Eastern Cape, including Port Elizabeth) City of Cape Town (in the Western Cape) The distribution and breakdown of local municipalities per province is as follows: Province Number of local municipalities Number of cross-boundary local municipalities Eastern Cape 38 Free State 20 Gauteng 7 1 with Mpumalanga 1 with North West KwaZulu-Natal 50 Mpumalanga 17 4 with Limpopo Province 1 with Gauteng North West 21 2 with Northern Cape 1 with Gauteng Northern Cape 24 2 with North West Limpopo 22 4 with Mpumalanga Province Western Cape 24 Total Table 1.1: Distribution of local municipalities across provinces Source: Demarcation Board Website ( From this it is clear that there are eight local municipalities that are located in two provinces. These are called cross-boundary local councils (CBLCs). These municipalities officially became operational in December 2000, but the transformation process did not occur at once. It comprised a gradual process of reorganisation, amalgamation, familiarisation with new processes and structures, etc. It also implied a reconsideration and redevelopment of many policy and strategy documents to include all the areas concerned. This was especially necessary since the demarcation involved the amalgamation of many smaller local councils (both urban and rural) into bigger municipalities. It also involved a lengthy process of re-assessing the new development CSIR Building and Construction Technology 7

14 challenges for the areas. One of the main challenges was and still is how to address crime at a local municipal level. CSIR Building and Construction Technology 8

15 Map 1: New Municipal Boundaries in South Africa CSIR Building and Construction Technology 9

16 1.6 Research methodology Objectives The purpose of the survey or national scan of gated communities was to determine the proliferation of gated communities and government responses in South Africa. The specific objectives were as follows: To determine the location of gated communities in SA To determine the extent of gated communities in SA To identify a typology of gated communities in SA To determine government s response to gated communities To use the information to identify relationships between different variables (for example, crime and income distribution and gated communities), and reach some conclusions on the impact and implications of gated communities in SA Research strategy, design and methods This suggested a quantitative research strategy, which can broadly be defined as one that emphasises quantification in the collection and analysis of data (Brymann 2001:20). The resultant strategy was expanded by making use of a cross-sectional research design 4 structured to cover the entire country in order to obtain a national perspective. The units of measure were all the metropolitan and local municipalities in South Africa (a total of 237). The first reason for this was that these municipalities would provide total spatial coverage of South Africa, and the second was that all matters such as local policies on types of gated community would fall within these municipalities and be applicable to these spatial areas. The method used to obtain the information about a range of variables to address the research objectives was self-complete questionnaires. The research methodology is summarised below: Research strategy Research design Research method Phase 1: Survey quantitative Cross-sectional Self-complete mail questionnaires Table 1.2: Research methodology for national scan The benefits of self-complete mail questionnaires are that they are less expensive and quicker to administer (especially given the target group of 237 municipalities), more accessible (allowing national coverage through the post), and they provide a similar and consistent base of data (comparable units and variables). The entire questionnaire was 4 A research design provides a framework for the collection and analysis of data and reflects the decisions about the priority being given to a range of dimensions of the research process, which may include causal relationships between variables, generalisation, etc (Bryman 2001:29). A crosssectional design entails the collection of data on one or more case (usually many more than one) at a single point in time in order to collect a body of quantitative data in connection with two or more variables which are then analysed to detect patterns of associations or relationships (Bryman 2001:41). CSIR Building and Construction Technology 10

17 made up of closed 5 questions to improve the response time, avoid any misunderstanding and provide a comparable base (see Appendix A). The first section dealt with the general presence of different types of gated community in the municipalities, based on a working typology developed during previous research (Landman 2000). These included enclosed neighbourhoods, large security estates, townhouse complexes and secured high-rise apartment blocks. This section also included questions regarding the consideration of the issue of gated communities within larger planning frameworks. Section B of the questionnaire was concerned with enclosed neighbourhoods, in particular, and Section C with large security estates. Again, previous research indicated that these two types are likely to have a greater impact on future urban planning and development, as well as on the urban form and fabric (Landman 2000). An initial set of 29 variables was identified in the questionnaire. The questionnaire was formatted into an A5 booklet to make it user-friendly, easy to read and follow, but used only one A4 page for the entire questionnaire. The questionnaires were numbered to allow diligent record-keeping and to keep track of the responding municipalities. All the questionnaires were accompanied by a covering letter and a stamped, addressed envelope to encourage easy return. The process also allowed for the option of faxing back the completed questionnaire. The covering letter (Appendix B) explained the purpose of the study, the process and the definitions of gated communities, enclosed neighbourhoods and security estates. The questionnaires were sent to all metropolitan (6) and local (231) municipalities in South Africa during February/March 2002, and the deadline for responses the end of July 2002, after which no responses would be accepted and those already received would have been analysed. While four months may seem long, it must be noted that it could take a long time for the questionnaires to reach the relevant local authority, and then be circulated to the department responsible for gated communities and the correct person working with these developments. Returned questionnaires confirmed that a number of different departments completed the questionnaire in different municipalities. This may be because gated communities are often the responsibility of more than one department in a local authority (for example, the planning and/or engineering and/or legal departments), and the corporate department often obtained the relevant information and completed the questionnaire itself. A month after the initial questionnaires where sent out, a follow-up request (including the questionnaire and covering letter) was faxed to all municipalities that had not yet responded Survey response The response rate was 49.4%, thus almost 50%, since 117 of the 237 municipalities had responded by the end of July The six metropolitan municipalities all responded. The responses are recorded per province (see Table 1.3), which distinguishes between those municipalities that fall completely within one province and those spanning more than one, namely the cross-boundary local municipalities (CBLCs), as mentioned above. The table also differentiates between the percentage of responses to the total number of questionnaires sent out, and that of municipalities that responded in a specific province. According to the total responses, the majority of respondents (21.37%) was from KwaZulu-Natal. This is to be expected, since this is also the province with the highest number of municipalities. The second-highest number of responses came from the Eastern Cape (10.26%), Limpopo (10.26%) and the North West (10.26%). 5 A closed question implies a closed answer, for example yes, no, or some, or important or unimportant, or a specific number. This is opposed to an open answer which may include someone s opinion, personal interpretation, etc. CSIR Building and Construction Technology 11

18 Table 1.3: Responses per province (metropolitan and local municipalities) Provinces Number of responses % of total response Number of municipalities % responding per province Eastern Cape Free State Gauteng GT/MP (CBLC) GT/NW (CBLC) KwaZulu-Natal Mpumalanga MP/LP (CBLC) Northern Cape NC/NW Limpopo North West Western Cape If one however, takes into account the fact that there are differing numbers of municipalities per province, a different picture emerges. Therefore, comparing the percentage response per province according to the number of municipalities per province indicates that, apart from the CBLCs, the highest percentage of response per province was from Gauteng (70%) and the Western Cape (68%). All nine provinces had almost a 30% response rate, with The lowest responses of all nine provinces were those of the Northern Cape (29%) and the Eastern Cape (30%). This is also summarised in the graphs below Percentages Eastern Cape Free State Gauteng GT/MP (CBLC) GT/NW (CBLC) KwaZulu Natal Mpumalanga MP/LP (CBLC) Northern Cape NC/NW Limpopo North West Western Cape Figure 1.7: Survey responses per province The following graph shows the relationship between the number of local municipalities in a province and the number of responses. One can clearly see that KwaZulu-Natal has the highest number of municipalities and therefore it is not surprising that it produced the highest CSIR Building and Construction Technology 12

19 number of responses. However, percentage-wise, only 49% of its municipalities responded. One reason for this may be that there is a large number of predominantly rural municipalities in the province that may not have an interest in what is mainly an urban phenomenon Percentages Number of responses 0 Number of municipalities Eastern Cape Free State Gauteng GT/MP (CBLC) GT/NW (CBLC) KwaZulu Natal Mpumalanga MP/LP (CBLC) Northern Cape NC/NW Limpopo North West Western Cape Figure 1.7: Survey responses per province in comparison with the number of municipalities per province The converse is that the high response rates in Gauteng and the Western Cape may have to do with the fact that these are the most urbanised provinces in South Africa, with Gauteng having an urban population of 97% and the Western Cape 89% (also see Figure 4.1 in Chapter 4). This inevitably raises the question of whether municipalities with a proportionately larger urban population, as well as municipalities with larger cities, would be more likely to respond to the questionnaire. When comparing the number of responses per category of percentage ***urban population in municipalities (Table 1.4), the highest percentage of responses (31%) came from municipalities with an urban population of 81 to 100%. The second-highest percentage response (29%) was however from the lowest category and not the second-highest category. If one compares the percentage response per category, i.e. the percentage of municipalities in relation to the number of municipalities in a particular category, the highest percentage response (62%) by far came from the category with the highest percentage urban population. The second-highest percentage response came from the middle category (52%) and the third highest (46%) from the second lowest category. It is therefore difficult to determine any significant relationships between the percentage urban population and the responses from different municipalities in the country. It is however to important to keep in mind that municipalities include both urban and rural areas, i.e. towns and farms or parks, as will be discussed in more detail later. Therefore a municipality may have several smaller towns with the same combined number of people than for example one large town. Therefore it may be that total percentage of urban/rural percentage may not be the best way to understand the responses. CSIR Building and Construction Technology 13

20 Table 1.4: Percentage urban population in municipalities and responses Percentage urban population Responses No of municipalities % response in category % No of % of total urbanised responses response 0-20 % % % % % Missing A more relevant way may be to look at the sizes of cities or towns within municipalities. It is likely that larger cities and towns will have higher numbers of enclosed neighbourhoods and large security estates. Those municipalities that responded included all the metropolitan areas and the majority of municipalities with larger cities and towns. Most of the municipalities that did not respond are predominantly rural municipalities with small towns. Their lack of response may be because different types of gated community are not a major concern in these areas, and perhaps not very prevalent. In this case, those that did respond may be a good representation of gated communities nationally. However, one still needs to avoid over-generalisation for all municipalities in South Africa Analysis of data The analysis of the questionnaires was done by looking for patterns of association or relationships between the initial 29 variables, as well as between them and an additional 6 other variables, to find possible reasons for the spread of gated communities in the country (see Appendix C for list of variables). The methods of analysis included univariate analysis through frequency tables (analysing one variable at a time), and bivariate analysis through contingency tables (to determine relationships between two variables at the same time) and Pearson s r (to determine correlations or relationships between interval/ratio variables). The statistical analysis was done by qualified statisticians from CSIR icomtek 6, using a computer programme (SAS) developed internally by one of their GIS specialists, Peter Schmitz. This was complemented by GIS mapping and categorisation, by both CSIR Boutek and icomtek, to express the relevant data spatially. One of the challenges was to find a way to compensate for the lack of correlation between the planning/administrative areas and South African Police Service (SAPS) boundaries, according to which the crime data are collected. CSIR icomtek (Peter Schmitz) developed a method to achieve this. The method comprises the union of the two layers for example, magisterial districts and police stations. The union creates new polygons, but keeps original attributes from both layers (in ArcView 3.2) (see Figure 1.8). The programme recalculates the areas of the new polygons in order to determine the proportionate relation of the new polygons to the original area (i.e. if after the union a police station is divided into three parts, then calculate the percentage area of each part in relation to the original area of the police station). The results are then exported into SAS. 6 CSIR Information and Communications Technology CSIR Building and Construction Technology 14

21 Figure 1.8: Creation of new polygons (union) from two different layers Figure 1.9: Crime data spatially displayed according to magisterial boundaries This method therefore provides a way to spatialise crime data according to planning regions, and provides a valuable way to understand spatial phenomena that are occurring as a response to crime and their implications for spatial planning in the city. CSIR Building and Construction Technology 15

22 Before one can discuss the distribution of the different types of gated communities, it is important to be clear about the definition and nature of the different types. The next section will be concerned with the typology of gated communities in South Africa. CSIR Building and Construction Technology 16

23 CHAPTER 2: CONFIRMATION OF TYPOLOGY 2.1 Working typology As mentioned before, previous research (Landman 2000) identified a working typology of gated communities in South Africa, which was taken as a starting point for this study. This typology comprises two main types of gated community, namely enclosed neighbourhoods and security villages. GATED COMMUNITIES IN SOUTH AFRICA Enclosed neighbourhoods Security villages Roads remain public property Roads become private property Large (mainly) residential security estates Secure office parks Residential security townhouse complexes Secure high-rise apartment blocks Figure 2.1: Definition and hierarchy of different types of gated communities in South Africa Enclosed neighbourhoods refer to existing neighbourhoods that have controlled access through gates or booms across existing roads. Many are fenced or walled off, as well, with a limited number of controlled entrances/exits and security guards at these points in some cases. The roads within these neighbourhoods were previously, or still are, in public ownership, depending on the model used. The previous survey (Landman 2000) indicated that there are different approaches to, or models of, enclosed neighbourhoods in South Africa within different municipalities. These include a public approach, a private approach, some combination of the two, or both. Municipalities may support one of the two, a CSIR Building and Construction Technology 16

24 combination of the two, or have both models from which residents can then choose. The implications of these two approaches are very different and are summarised in the diagram below: Approaches to enclosed neighbourhoods Public Private Public spaces owned and maintained by the local authority Private spaces owned and maintained by the homeowners' association Cannot legally restrict access into these areas, because spaces inside are still public spaces Can legally restrict access into these areas, because the spaces inside have become private spaces Figure 2.2: Public and private approaches to enclosed neighbourhoods If the roads, parks, sidewalks, etc, are still owned by the local authority, the local authority is responsible for the maintenance of these areas (public approach). If the areas have, however, been taken over by the residents association, these areas become private space and the residents are responsible for their maintenance (private approach). Enclosed neighbourhoods also have different implications for accessibility. According to the South African Constitution it is the right of all people to have access and free movement to all public space. 1 The important issue then is whether the enclosed area remains under public control or is taken over as private space by the residents/homeowners association. If the enclosed area stays under public control, all people have the right to enter the public spaces within this area, and provision should be made for them to be able to do so at all times. 1 Section 21 (1) and (3) of the Constitution of the RSA, 1996 (Act 108 0f 1996) reads as follows: (1) Everyone has the right of freedom of movement (3) Every citizen has the right to enter, to remain in and to reside anywhere in the Republic. CSIR Building and Construction Technology 17

25 Security Villages 2 refer to developments where the entire area has been developed by a private developer. These areas/buildings are physically walled or fenced off and usually have a security gate or controlled-access point with/without a security guard. The roads in these developments are private and, in most of the cases, their management and maintenance is carried out by a private management body. Security villages include not only residential areas (such as townhouse complexes, high-rise apartment blocks), but also controlled-access villages for business purposes (office blocks) and mixed-use developments such as large security estates. Although many of these are predominantly residential, a growing percentage of other land uses are found within these developments, including commercial (golf shops, post offices or boxes, estates agents, etc) and/or recreational (golf courses, squash courts, tennis courts, equestrian routes, etc) uses. Although this working typology includes both enclosed neighbourhoods and security villages, the previous research focused mainly on enclosed neighbourhoods. This was due to the limitations of time and resources, as well as the need to look into enclosed neighbourhoods which were emerging as a significant phenomenon at that stage (both in terms of number and the responsibilities of, or responses from, local municipalities). The research however found that there is a need for more in-depth research into this subject in South Africa, and that looking at only part of the subject will not be sufficient. Enclosed neighbourhoods are only one part of the bigger phenomenon of gated communities and any attempt to understand the entirety of the phenomenon in South Africa will therefore have to look at both of the main types both enclosed neighbourhoods and security villages, since both are likely to have significant but probably different impacts on urban planning and development in the future. 2.2 Confirmation of working typology This study used the working typology as a basis for further exploration. The first section of the questionnaire aimed to determine the existence and spread of different types of gated community in the country. The results are summarised in the table below: Table 2.1: Occurrence of types and sub-types of gated communities for all municipalities Encl neighb. Security villages Enclosed neighbourhoods Office parks Security estates Secure townhouses Secure highrises N % N % N % N % N % Yes No Don't know Missing / Not completed About one-fifth of all respondents indicated that they had enclosed neighbourhoods in their municipalities. The survey also confirmed the existence of the four types of security village identified in the working typology. Only a small percentage had office parks (6%). This may 2 The term village can refer to anything from a small collection of houses in a rural area, to a large collection of houses in an urban area. The last mentioned have often been referred to as an urban village. In the case of security villages, it refers to a variety of developments, ranging from a small collection of buildings (for office use) or houses, to a large collection of houses etc, inside a protected boundary (through a fence or wall) and entrance/s with controlled access. CSIR Building and Construction Technology 18

26 be due to the fact that secure office parks tend to occur only in larger cities. Larger percentages of high-rise apartment blocks (20%) and security estates (24%) were recorded. Both of these also tend to occur in the larger cities. The percentage of townhouse complexes (35%) was considerably higher than all the other types (also see Figure 2.3). One explanation may be that they tend to occur not only in larger cities, but also in medium-sized cities and larger towns. In addition, this terminology is better known, and may have influenced the responses. Percentage of all respondents saying "yes" to existence of different types Encl neighb Office parks Security estates Secure townhouses Secure high rises Figure 2.3: Percentages of different types of gated community in South Africa (including only those municipalities who responded) From the number of local authorities who responded as having more than one type, the survey showed that 57% of respondents have existing enclosed neighbourhoods and new security office parks (See Appendix E). This is very interesting, since only 6% of the overall number of respondents had office parks. It may suggest that both these types of very prevalent in larger cities. Compared to this, 53% of the local authorities had both enclosed neighbourhoods and security estates; almost matching the previous group. Therefore, more than half of the responding municipalities had both enclosed neighbourhoods and large security estates. However, when one compares the number of local authorities with both enclosed neighbourhoods and secure townhouse complexes, the percentage is much lower (36%). This is interesting when one considers that the type that occurred most in the local authorities that responded (35%), was secure townhouses (see Figure 2.3). The percentage of local authorities with both enclosed neighbourhoods and secure high-rises (39%) is also lower than the first two (those with enclosed neighbourhoods and security estates or office parks) but slightly higher than the previous group (Appendix E). One reason for the lower percentages in these categories may be that a third option was included for the last two types in the questionnaire, namely don t know. A number of municipalities chose this option (see Table 2.1). Another reason for the lower percentage in the first category, which will be explored in more detail later in the report, may be that enclosed CSIR Building and Construction Technology 19

27 neighbourhoods are a more recent physical response to crime and may be occurring more in larger cities. Because they are a more recent phenomenon, they may not yet be a popular way of responding in smaller cities and towns, where smaller townhouse complexes may be a more popular choice and thus more prevalent. This study, however, focuses only on two types, which were investigated in more detail. They were chosen from both categories thus one type of enclosed neighbourhood and one type of security village, with specific consideration of the likely impact on, and implications for, urban transformation and future planning and development, as well as decision-making and policy formulation. It was decided to focus on enclosed neighbourhoods, which are based on a public approach thus where the urban spaces (streets, sidewalks, parks, etc) still remain public property and are thus classified as public space or the public realm. In this case the local authority is still responsible for the maintenance and upgrading of the public space. These types of development have major implications for urban management, the use of urban space, the transformation of urban space and the nature of the public realm. The other type that the study focuses on is large, mainly residential, but often mixed-used security estates. These types, especially the larger security estates, are likely to have a bigger impact on the urban form and fabric (due to their size), as well as on the lifestyle and perceptions of urban residents (due to the regulations and controls within these areas). The questionnaire was thus structured to focus on these two types in more detail. The questionnaire did not explicitly distinguish between the two models of enclosed neighbourhood. There were two reasons in particular for this approach. It was felt that this would confuse the target group (since a large number was not familiar with the broader term, let alone the distinction) and the previous research showed that very few local municipalities, in fact, allow and encourage a private response. It was also interesting that those who allowed for both options had almost none that opted for the private approach. This could be due to the high maintenance and upgrading costs involved in the privatisation of public space and infrastructure. The second phase of the research, namely the case studies, however included only enclosed neighbourhoods based on the public approach. The questionnaires did however clearly differentiate large security estates, while the case studies also included two cases from this category. The remainder of the discussion and analysis will therefore be concerned with only these two types, bearing in mind that in the survey and discussion enclosed neighbourhoods refers to both sub-types (although most are based on a public approach) and security estates to large mainly residential security villages. 2.3 Clarification of terminology regarding the typology and terms used in the questionnaire and survey analysis The questionnaire referred to the different typologies mentioned above, namely enclosed neighbourhoods, large (mainly residential) security villages/estates, office parks, secure townhouse complexes and secure high-rise apartment blocks. Enclosed neighbourhoods were defined as existing neighbourhoods that have been fenced in through the closure of public roads. Large security estates were defined as large, luxury security villages or estates, such as golfing or country estates, and other large mainly residential security areas. It was also pointed out that smaller security townhouse complexes - with, fewer than 50 units were not included in this category or type. However, it may have been confusing that some questions referred to enclosed neighbourhoods and others to road closures. Simply put, road closures refer to the physical closure of a particular road through a boom or gate or other physical structure. An CSIR Building and Construction Technology 20

28 enclosed neighbourhood refers to an entire area that has been enclosed through one or more street closures that is, controlling access to the entire neighbourhood or area. The sizes of neighbourhood enclosures differ considerably and often range from an area controlled by only one road closure to areas that have up to 23 road closures controlling access into one neighbourhood (see Figure 2.4). Figure 2.4: Distinguishing between road closures and an enclosed neighbourhood. The thick black line on the left indicates the enclosed neighbourhood (example from Johannesburg) and the enlargement on the right indicates the actual road closures (total of 23 in this case) constituting the enclosed neighbourhood. The black dots indicate accessible road closures (i.e. through gates or booms that can be opened to allow traffic to pass) and the red crosses indicate inaccessible road closures (through fences and gates that are locked). Applications and approvals for enclosed neighbourhoods are submitted and approved per neighbourhood, regardless of the number of roads inside that have to be closed. Therefore questions 15 and 16 refer to enclosed neighbourhoods in the collective sense a collection of road closures constituting an enclosed neighbourhood. It is also important to distinguish between temporary and permanent road closures. Roads may be closed by the traffic department from time to time to manage traffic flow, etc through curbs, no entry signs, etc. Roads that are closed for security to allow security access restriction may only do so on a temporary basis for two years. Although they also constitute permanent physical structures (for example brick guard houses, fences and gates), these are technically a temporary road closure. A security estate, refers to the entire enclosed area as an entity, regardless of how many entrances there are into the area. CSIR Building and Construction Technology 21

29 CHAPTER 3: DISTRIBUTION OF GATED COMMUNITIES IN SOUTH AFRICA 3.1 Overall distribution in provinces As already mentioned, a total number of 117 municipalities responded to the survey. Of these, 23 (20%) confirmed the existence of neighbourhood enclosures, while 28 (23%) said they had large security estates. As mentioned in Chapter 1, it is assumed that there are limited numbers of enclosed neighbourhoods and security estates that occur within those that did not respond. Missing Yes 1% 20% Missing Yes 20% 23% No 79% Figure 3.1: Percentage of local authorities with enclosed neighbourhoods No 57% Figure 3.2: Percentage of local authorities with security estates Although the percentages of local municipalities with enclosed neighbourhoods (20%) and security estates (23%) are more or less the same (according to those who responded), the distribution of these two types per province differs substantially. Gauteng (30% of total response) and the Limpopo Province (17%) had the highest number of local municipalities with enclosed neighbourhoods, followed closely by the Western Cape (13%). On the other hand, KwaZulu-Natal and the Western Cape had the highest number of local authorities with security estates, followed by Gauteng and Limpopo Province. It must however be kept in mind that the number of local municipalities per province also differs substantially, as was pointed out in Chapter 1. This may have an influence on the number of local authorities with higher numbers of enclosed neighbourhoods or security estates. For example, KwaZulu-Natal has the highest number of local municipalities and also the highest number of security estates. On the other hand, Gauteng has the lowest number of municipalities but the highest number with enclosed neighbourhoods, and the secondhighest number with security estates, if one only considers those local authorities that indicated that the phenomenon existed within their jurisdictional areas (Appendix F). One province, Free State, reported no local authorities with enclosed neighbourhoods, but had one with security estates. This may be due to the fact that enclosed neighbourhoods are not permitted in the largest municipalities with the largest cities in the province (Bloemfontein and Bethlehem). Although the Dihlabeng Municipality (including Bethlehem) allowed the establishment of enclosed neighbourhoods in the past, this practice was ruled unconstitutional at a recent council meeting and the few existing road and neighbourhoods closures had to be removed. The opposite is true in another province, Northern Cape, which reported no local authorities with security estates, but has one with enclosed neighbourhoods CSIR Building and Construction Technology 22

30 (Appendix F). KwaZulu-Natal reported only one local authority as having enclosed neighbourhoods, while six (24% of municipalities in the province and 21% of the total) indicated the existence of security estates. One explanation may be that many municipalities in this province do not allow the establishment of enclosed neighbourhoods for example, e Thikweni metropolitan municipality (including Durban). The result has been either to limit their occurrence, or illegal closures, as in the case of e Thikweni. Three provinces, Gauteng, Limpopo Province and the Western Cape have the highest numbers of local authorities with both enclosed neighbourhoods and security estates. Gauteng had the highest percentages of municipalities per province with enclosed neighbourhoods (77%) and security estates (44%), in terms of the total number of provinces. One reason for this may be the high response rate (77% of all municipalities responded) in Gauteng, but there may also be a range of other reasons, which will be further explored in the next chapter. When comparing the distribution of these two types per province, two things must always be kept in mind. First, the comparison between provinces as discussed above did not include the CBLCs, because the low numbers skew the data. Secondly, when one considers the proportion of yes responses in the province in the case of security estates, it should be kept in mind that the missing number was 23 (not completed or considered not applicable), which leaves a large proportion (19%) unaccounted for Enclosed neighbourhoods Security estates 0 Eastern Cape Free State Gauteng GT/MP (CBLC) GT/NW (CBLC) KwaZulu Natal Mpumalanga MP/LP (CBLC) Northern Cape NC/NW Limpopo North West Western Cape Figure 3.3: Number of local authorities with enclosed neighbourhoods and security villages per province (including only those who responded) 3.2 Distribution of different types in municipal areas in South Africa The numbers of enclosed neighbourhoods and security villages within municipalities differ substantially. This is true both for the number of one type in different municipalities, and for the distribution of both types in municipal areas. The survey indicated the highest numbers of enclosed neighbourhoods in Gauteng, with two municipalities having from seven to nine neighbourhood closures, two having from 16 to 25, one from 25 to 100 and one with more than a hundred (see Maps 2 and 3). Two of the three metropolitan municipalities had the highest numbers of enclosed neighbourhoods at the time of the response namely Tshwane with 36, and Johannesburg with 300. There were also two municipalities in the Western Cape with higher numbers, namely those that included Cape Town and Mossel Bay. The City of Cape Town metropolitan municipality recorded 25 neighbourhood closures at the time of response, and the Mossel Bay municipality 20. CSIR Building and Construction Technology 23

31 Map 2: The distribution of road closures in South Africa CSIR Building and Construction Technology 24

32 Map 3: The distribution of enclosed neighbourhoods in Gauteng CSIR Building and Construction Technology 25

33 There are also a number of illegal closures in many municipalities. Five municipalities were aware of illegal neighbourhood closures in their areas. There are reported to be about 500 illegal road closures in the City of Johannesburg at present. The distribution of security estates differs to some extent. While some of the municipalities mentioned reported high numbers of security estates, such as the cities of Johannesburg (20), Tshwane (18) and Cape Town (24), high numbers were reported in other municipalities that recorded relatively low numbers of road closures or none at all (see Map 4). The highest numbers of security estates were recorded in Emfuleni (40) and Madibeng (31) municipalities. Emfuleni is located in the south of Gauteng, including cities such as Vanderbijlpark and Vereeniging, and Madibeng is located in the south-east of the North West Province on the Gauteng boundary, and includes towns such as Brits and Hartbeespoort (see Map 5). Other municipalities with high numbers of security estates include Plettenburg Bay (21) and Knysna (20) in the Western Cape, 3.3 Distribution of different types within cities The questionnaire included two questions regarding the location of enclosed neighbourhoods and security estates within cities in the different municipalities. The highest numbers per municipality, for both types, occurred in peripheral suburbs (see Table 3.3). In the case of enclosed neighbourhoods (Figure 3.4 below), 38% of the municipalities said that they tended to occur in peripheral suburbs, 27% reported them in inner suburbs and 23% indicated their presence in more than one area (mostly peripheral suburbs and inner suburbs). There is not such a large difference between those occurring in peripheral suburbs and inner suburbs, suggesting that enclosed neighbourhoods are more or less spread equally between these larger urban areas. What is, however, of particular interest is that one municipality (4% of the total) indicated that it had enclosed neighbourhoods in traditional townships. Compared to this (see Figure 3.5), 53% of the municipalities indicated that security villages tended to occur in peripheral suburbs, 20% reported them in inner suburbs and 27% in more than one area (mostly peripheral suburbs and inner suburbs). This shows a clear predominance of large estates in peripheral suburbs. One reason could be that such developments usually require large greenfield sites that are mostly available only on the urban periphery 1. Enclosed neighbourhoods and security estates both occur in peripheral suburbs. This suggests that these two types are popular ways to increase safety through physical intervention in the newer areas in the city. 1 The percentages in both of these groups only include rows one to six of Table 3.3, thus exclude the municipalities that did not complete these questions because they were not relevant to them (missing data or N/A). CSIR Building and Construction Technology 26

34 Map 4: The distribution of large security villages/ estates in South Africa CSIR Building and Construction Technology 27

35 Map 5: The distribution of large security villages/ estates in Gauteng CSIR Building and Construction Technology 28

36 Table 3.2: Location of enclosed neighbourhoods and security estates within cities in municipalities Locations in cities in municipalities Enclosed neighbourhoods Security estates No of municipalities No of municipalities Urban core 1 0 Inner suburbs 7 6 Peripheral suburbs Traditional townships 1 0 More than one 6 8 Rural 1 0 Missing or N/A More than one 23% Enclosed neighbourhoods Rural 4% Urban core 4% Inner suburbs 27% Figure 3.4: The proportion of enclosed neighbourhoods in different parts of the city Traditional townships 4% Peripheral suburbs 38% Security estates Rural 0% Figure 3.5: The proportion of security estates in different parts of the city More than one 27% Urban core 0% Inner suburbs 20% Traditional townships 0% Peripheral suburbs 53% Given the distribution of these two types of gated community in the country in different provinces and municipalities, as well as within cities questions that inevitably come to mind that can be addressed include why they are located in these geographical areas, what the factors influencing the distribution of these types of development are, and whether there are dominant influences giving rise to particularly high numbers in certain areas. These questions will be explored in the next chapter. CSIR Building and Construction Technology 29

37 CHAPTER 4: POSSIBLE REASONS OR RELATIONSHIPS FOR THIS DISTRIBUTION Chapter 1 indicated that gated communities in South Africa are generally regarded as a physical response to crime and the fear of crime. The first part of this chapter (Section 4.1) will therefore investigate whether there is a link between the responses to the survey, in terms of the location of enclosed neighbourhoods and security estates and the distribution of crime in South Africa, in terms both of provinces and municipalities. It will also explore this relationship within two cities, namely Johannesburg and Tshwane. The survey indicated that, according to those who responded, Gauteng had the highest number of municipalities with both enclosed neighbourhoods and security estates, as well as very high numbers of these developments. Although these figures were not corrected for population, they are likely to be an issue within these municipalities, which at best may have to deal with the phenomenon. The other major driver that may influence the distribution and numbers of enclosed neighbourhoods may be income. Municipalities charge application fees, some of which are quite high, to process an application for a neighbourhood closure. In addition, there are often other costs involved, such as private security, garden services, financial auditing, management fees, etc. Communities must therefore have access to additional funding (relatively higher disposable incomes) to be able to embark on such an action. Property prices, as well as monthly levies within large security estates are also very high. A recent article have shown that properties in many of these estates in Gauteng start at around R1.5 million and can go up to R10 million (Fife 2002). The second section (4.2) will investigate whether there is a link between the survey responses, in terms of the location of enclosed neighbourhoods and security estates and the distribution of income in South Africa, both in terms of provinces and municipalities. Although these may be the predominant drivers, there are other factors that could play a role, such as the type of municipality (metro, rural, etc), the size of municipality (according to population and square kilometres), the size of cities within a municipality and whether some of the cities within the municipality are located closed to natural amenities, such as the sea, a major dam, etc. These relationships will be explored in Section Relationship of gated communities with high crime areas Previous research in South Africa has identified crime as the major reason for gated communities in the country (Bremner 1999; Landman 2000, Vrodljak 2002; Van Wateren 2002). The aim of this section is to determine whether there is any relationship between crime in provinces and the number of enclosed neighbourhoods and security estates in those provinces, as well as between crime in the municipal areas and the number of both types in these areas. Crime patterns are indicated spatially according to provinces and municipalities, and in terms of overall crime rates, property crimes and violent crime. As mentioned earlier, the majority of crimes that occur in South Africa are property crimes. This is the pattern worldwide. What is of more concern is the unacceptably high level of violent crimes in the country (Schönteich 2001). CSIR Building and Construction Technology 30

38 Table 4.1: Types of violent and property crime (as used for categorisation on all the crime maps) Violent crimes Property crimes Murder Burglary - business Attempted murder Burglary - residential Robbery 1 : aggravated - firearm Theft of motor vehicle Robbery: aggravated - no firearm Theft out of/from motor vehicle Other robbery Other theft Child abuse Arson Kidnapping 0-17 yrs Malicious damage to property Kidnapping 18+ yrs Assault, serious Assault, common Carjacking Truckjacking Robbery - cash in transit Bank robbery Crime and gated communities in provinces The highest total crime rates per province for the period 1998 to 2001 were found in Gauteng, the Western Cape and the Northern Cape (Map 7). The Limpopo Province had the lowest crime rates. Two of the three provinces with the highest crime rates also have high numbers of municipalities with enclosed neighbourhoods and security villages namely Gauteng and the Western Cape. On the other end, however, the province with the lowest crime rate also had high numbers of enclosed neighbourhoods and security estates. One possible explanation for this pattern could be the percentage response rates in the Northern Cape and Limpopo provinces. While Limpopo Province had a relatively high response rate per province (54.55%), the Northern Cape recorded the lowest response rate of all the provinces (29.17%). Correlations 2 between the number of municipalities per province with enclosed neighbourhoods and large security estates revealed different findings for each of the five years, and for the average for the five years. Only the Western Cape showed a significant correlation (p-value = ) between the number of municipalities with enclosed neighbourhoods and the average crime rate for 1998 to 2001 (see Appendix F). In addition, only the Eastern Cape and KwaZulu-Natal showed significant correlations (p-values = and < respectively) between the number of municipalities with security estates and the average crime rate between 1998 and 2001 in these provinces. This could mean that there is a strong relationship between crime and the number of municipalities with these developments in these provinces. 1 Robbery with aggravating circumstances and other robbery would include all crimes where property is taken from a person by force. Although robbery is therefore technically a violent crime since it involves the use of force, in South Africa it is often put into a separate category because of its seriousness and because it is associated with the theft of property. The dynamics associated with assault, rape and murder are also different from those linked with robberies. 2 All the correlations in this document made use of Pearson s Correlation Coefficients (Pearson s r ) to determine whether there are any significant correlations between variables. 3 The p-value refers to the probability value. Any correlation with a p-value of <0.05 is considered statistically significant. A p-value of <0.05 implies that there is a 95% or more chance that the values are indeed correlated. CSIR Building and Construction Technology 31

39 Map 7: Total crime rates per province for the period CSIR Building and Construction Technology 32

40 However, it is important to note that this does not give an indication of numbers of these developments in the provinces. It also does not take into account variables such as the size of urban agglomerations, densities, etc, which can all have an impact on crime, as well as the fear of crime. It may also be that the use of absolute numbers influences the correlation. Finally, the number of the responses ( n value) is very small in some cases, due to the fact that some provinces have more municipalities than others, and others also have more of the two types of developments under consideration. In addition, many provinces did not reply to or complete these particular questions and these missing values were excluded from the correlation. This resulted in very few municipalities being taken into account in some provinces. Some provinces had an n value of less than 10, especially those with very few municipalities or none of these types of developments. In these cases, they cannot be used. The results also showed that most of the provinces with a very small n value did not show any significant correlations. This may however be not because the n value was too small, but because there were not enough values to confirm the significance of the correlation. Therefore, this may not be an appropriate method to test for any relationship in the case of many of the provinces Another explanation could have to do with the types of crime that are most prevalent in these provinces. Although the categories (range of numbers of incidences) differ slightly, the distribution of violent crime in the provinces follows the same pattern as the total crime rates in terms of the provinces with the highest rates Gauteng, the Western Cape and the Northern Cape (see Map 8). There is, however, a different pattern in terms of the lowest crime rates, where four of the provinces, namely Limpopo Province, North West, KwaZulu- Natal and the Eastern Cape, fall within the lowest category, thus having less than crimes per 1 million persons. The picture changes somewhat when comparing the property crime rates per province (Map 9). The Limpopo Province still falls within the lowest category, but this category now includes < property crimes compared to the < violent crimes. On the other side of the spectrum, Gauteng and the Western Cape have the highest number of property crimes (> per 1 million population), followed by the Northern Cape ( ). In this case, Gauteng and the Western Cape therefore clearly stand out as the provinces with the highest property crime rates from 1998 to Both of these provinces had high numbers of security estates and enclosed neighbourhoods in some municipalities, as indicated in the previous chapter. It is however true that the distributions of types of crime within municipalities and cities differ. Studies have shown that property crimes are more prevalent in wealthier suburbs and violent crimes in poorer, traditional township areas (Shaw and Louw 1998). This can make it very difficult to generalise, but it would suggest a stronger relationship between areas with higher property crime and both enclosed neighbourhoods and security villages. Yet it is difficult to measure this on a provincial scale due to additional factors that could influence the outcome for example, the fear of crime Fear of crime and gated communities in provinces and municipalities As discussed in Chapter 1, the HSRC survey (1999) revealed that the highest percentage of people feeling unsafe were from the Eastern Cape (57%) and Gauteng (51%). Gauteng also had the lowest number of people feeling safe (34%). When comparing this to the number of municipalities with enclosed neighbourhoods and security estates, there appears to be a relationship between the high levels of fear of crime in Gauteng and the high number of municipalities with these types of gated community. The fear of crime in the Eastern Cape, however, did not appear to have resulted in a large number of gated communities in terms of number of municipalities reporting their existence. Therefore one may also have to correct for the wealthier proportion of the population per province, since these two types of responses may be an option only for people with higher incomes. CSIR Building and Construction Technology 33

41 Map 8: Violent crime rates per province for the period CSIR Building and Construction Technology 34

42 Map 9: Property crime rates per province for the period CSIR Building and Construction Technology 35

43 Crime and gated communities in municipalities Map 10 (below) spatialises the overall crime rates per municipal area for the period 1998 to 2001 according to five broad categories. Only two municipalities fall within the highest category (> crimes per population) and both of these are located in the Northern Cape. The one to the west (rendered in dark brown), includes the town of Upington and is known as the Il Khana Hais municipality. The one on the provincial boundary with the Free State to the east includes the city of Kimberley, and is known as the Sol Plaatjie local municipality. Kimberley is also the capital of the Northern Cape. Neither of these municipalities has enclosed neighbourhoods or security estates. One reason may be that violent crime predominates in the area and that enclosed neighbourhoods and security estates are more a response to property crime. International and local research on environmental design to prevent crime has often indicated that environmental design is often more effective in addressing property crime. 4 While environmental design can reduce opportunities for violent crime such as murder and rape in public spaces, there is very little that it can do to address domestic personal violence (Shaw and Louw 1998). When comparing the distribution of violent crime in municipalities (Map 11), both these municipalities indeed show the highest violent crime rates (>550) per persons. They even have higher violent crime rates than the metropolitan areas. If one, however, considers the distribution of property crimes per municipality (Map 6), the picture changes completely. In this case, there are 29 municipalities in the highest category (>550). Of these, 34.5% are located in the Western Cape, followed by 17% in the Northern Cape. It is also noteworthy that the Free State had no municipalities in the highest category, while the Limpopo and Eastern Cape had only one each. Of these 29 municipalities, 20 responded (69%) to the questionnaire. When one compares these municipalities falling in the highest category for property crimes with the existence of either enclosed neighbourhoods or security estates, 38% reported the existence of enclosed neighbourhoods, compared to almost half (48%) which had large security estates. In the case of enclosed neighbourhoods, 11%, however, indicated that they had received applications for enclosed neighbourhoods. Almost one third of the municipalities in this category did not respond (Figures 4.4 and 4.5). DNR 31% Yes 38% DNR 31% Yes 48% No (applic) 10% No 21% No 21% Figure 4.4: Existence of enclosed neighbourhoods in high property crime municipalities Figure 4.5 Existence of large security estates in high property crime municipalities 4 A number of authors consider enclosed neighbourhoods and security estates as an effective example of Crime Prevention through Environmental Design (CPTED) (Newman 1995; Foldarvry 1994), while others acknowledge it as CPTED, but warn of the often negative consequences that result from such actions (Blakely and Snyder 1997; Ellin 1997). CSIR Building and Construction Technology 36

44 Map 10: Total crime rates per municipality for the period CSIR Building and Construction Technology 37

45 Map 11: Violent crime rates per municipality for the period CSIR Building and Construction Technology 38

46 Map 12: Property crime rates per municipality for the period CSIR Building and Construction Technology 39

47 The overall crime rates in the municipalities and the number of security villages showed a significant correlation for all four years, and the average crime rates from 1998 to 2001 did as well (see Appendix G). There was however also a significant negative correlation between the number of municipalities with enclosed neighbourhoods and security estates and the crime in municipalities in all categories. Again, the problem may be the use of absolute numbers. In the case of violent crimes in municipalities there was a positive relationship between violent crimes and the number of security estates in all categories, and a negative relationship between violent crimes and municipalities with enclosed neighbourhoods. On the other hand, there was a significant correlation (positive relationship) between property crimes and the number of security villages in municipalities in all the categories, as well as between property crimes and municipalities with enclosed neighbourhoods (Appendix G). Again, there may be a problem with the use of absolute numbers. In terms of fear of crime, the HSRC survey (1999) indicated that the lowest percentage of respondents feeling safe was from metropolitan areas (35%), followed by urban areas (43%). The highest percentage of people feeling safe was in rural areas (53%). When this is compared to the number of municipalities with enclosed neighbourhoods, it appears that the feeling of safety may have an influence on the occurrence of enclosed neighbourhoods and security estates, although it is not possible to say too what degree. According to the responses, four of the six metropolitan municipalities have both enclosed neighbourhoods 5 and security estates Crime and gated communities in the cities of Johannesburg and Tshwane The discussion up to now has revealed that the highest number of enclosed neighbourhoods and security villages tend to occur in municipalities with a proportionately large urban population, or with medium to large cities (including metropolitan areas). It has also shown that both of these types tend to occur in municipalities with high property crime rates. The study will therefore focus in much more detail on two municipalities with all these characteristics. They are the City of Johannesburg and the City of Tshwane. Both of these municipalities have proportionately very high urban populations and are two of the largest urban areas in the country with populations of over one million (Table 4.10). As municipalities they also both fall within the highest category of property crime rates in the country (Map 11). Table 4.2: Description of selected information in the cities of Johannesburg and Tshwane Property Number of crime rates road Population (2001) % Urban population ( ) closures City of Johannesburg City of Tshwane Overall crime rates ( ) per persons per persons >550 per persons >550 per persons Number of security estates In the survey e Thekweni (including Durban) responded no, because it does not allow them at present. It does, however, have numerous illegal neighbourhood closures. The Nelson Mandela metropole (including Port Elizabeth) also responded no, but in a previous survey indicated the existence of enclosed neighbourhoods, although in its case it only allows them if the residents completely take over the area concerned and manage and maintain it as a private area; this is the private approach to enclosed neighbourhoods. 6 e Thekweni also has large security estates. CSIR Building and Construction Technology 40

48 City-wise (per police areas), the two also have the highest overall crime rates in South Africa, compared to other cities that fall within metropolitan areas. Both of these municipalities also indicated the highest numbers of road closures in the country and among the highest numbers of security estates. Recorded crimes in selected police areas, 2000 Figure 4.6: Recorded crime in selected police areas, 2000 (Shönteich 2002b:3) Durban 6115 Port Elizabeth West Metrople (CT) Pretoria Johannesburg No. per population Following on from the discussion in the previous section, one of the areas of investigation will be whether there is any relationship between enclosed neighbourhoods and security estates in terms of property crimes. Schönteich and Louw (2001) point out that there are particular patterns in South Africa pertaining to the occurrence of violent crime and property crime in certain parts of the city. Often, the incidence of property crime in a certain area is frequently inversely related to the incidence of violent crime. Townships and poorer areas, which experience the highest per capita levels of violent crime in cities, often also have the lowest levels of property crime. On the other hand, wealthier suburbs tend to experience exactly the opposite. City centres or inner-city areas often experience high levels of both property and violent crime (Schönteich and Louw 2001: 6). This suggests a strong relationship between the average income in an area and the prevalence of certain types of crime. Again, there is a lack of correlation between police-station areas/boundaries and planning and administrative areas/boundaries (according to the new municipal areas). In order to overcome this problem, the crime data was adapted to the new IDP areas by CSIR icomtek, as discussed in Chapter 1. a) Crime and gated communities in the City of Johannesburg Although Johannesburg has the highest crime rates in the country, crime is not equally distributed throughout the municipal area. According to the Johannesburg IDP the municipal area is divided into 11 planning/administrative regions (see Appendix F for population distribution per region from ). In terms of overall crime rates, the Central Region has the highest crime rate (>27 000) per persons for the period 1998 to 2001 (Map 13 below). Sandton/Rosebank (Region 3) has the second highest crime rate ( ) per persons in the municipal area. Four areas fall within the middle category ( ), namely Midrand/Ivory Park, Alexandra, Northcliff and Ennerdale/Orange Farm. CSIR Building and Construction Technology 41

49 The distribution of violent crime rates, however, differs completely (Map 8). While the central area still has the highest rate (>1800 per persons), all the other areas have much lower rates in comparison, with three falling in the second-lowest category and all the rest in the lowest category. The distribution of property crimes (Map 14), is much more reflective of the distribution of the overall crime rate (Map 15). Category-wise, although the actual numbers in the categories differ, the regions all fall within the same categories except Ennerdale/Orange Farm. As with the overall crime rate, the Central Region has the highest property crime rates (>1800 per persons), followed by Sandton/Rosebank ( per persons). In this case, three regions fall within the middle category ( property crimes per persons) namely Midrand/Ivory Park, Alexandra and Northcliff. If one assumes that the same pattern exists within cities as in the case in municipalities (i.e. that there is some sort of relationship between the existence of enclosed neighbourhoods/security estates and high property crime rates), one would expect these types of gated communities to be located in the areas with the highest numbers of property crimes. This appears to be the case in Johannesburg to some extent, although there are other factors that influence the distribution and type of crime. According to its returned questionnaire, the City of Johannesburg indicated that the majority of enclosed neighbourhoods occur in Randburg, Sandton, Alberton (in Johannesburg South), Midrand and Roodepoort (Question 9). Sandton fell within the second-highest category of property crimes, while Randburg and Roodepoort (part of Roodepoort region) and Midrand fell in the middle category. Alberton (part of Johannesburg South) fell in the second-lowest category. On the other hand, there was no indication of high numbers of enclosed neighbourhoods in the Central Region, which had the highest property crime rates. This can be partly explained by two factors. Firstly, there are other types of gated communities, and enclosed neighbourhoods are not always the most appropriate form of response since they often require substantial resources both for development and operation. The City of Johannesburg also indicated the existence of secure office parks, secure townhouse complexes and secure high-rise apartment complexes in its response to the questionnaire. Secure high-rise apartment complexes are especially prevalent in the Central Region, which may also relate to the specific housing needs in this area: a greater need for lower-middle to low-income housing options. This may also suggest a possible relationship between enclosed neighbourhoods and middle to higher income areas, which will be investigated later in this chapter. The second factor influencing the distribution may therefore be the average income per area. According to the questionnaire, security estates were the most prevalent in the same area, namely Randburg, Sandton, Alberton (part of Johannesburg South), Midrand and Roodepoort (Question 19). Again the explanation could be the same as in the case of enclosed neighbourhoods, with type of gated community and average income playing an important role, in addition to crime, in terms of the distribution of security estates in Johannesburg. Apart from the crime rates and average income, certain types of gated communities may be more likely to occur in certain urban areas. Therefore, while secure high-rise apartment blocks may be more prevalent in inner-city areas, enclosed neighbourhoods and security estates may be more prevalent in suburban areas. The response from the City of Johannesburg indicated that enclosed neighbourhoods and security estates both occur mainly in the urban periphery. CSIR Building and Construction Technology 42

50 Map 13: Total crime rates per persons for the period in the City of Johannesburg CSIR Building and Construction Technology 43

51 Map 14: Violent crime rates per persons for the period in the City of Johannesburg CSIR Building and Construction Technology 44

52 Map 15: Property crime rates per persons for the period 1998 to 2001 in the City of Johannesburg CSIR Building and Construction Technology 45

53 b) Crime and gated communities in the City of Tshwane As in the case of Johannesburg, crime is not uniformly distributed in the City of Tshwane. According to the Tshwane IDP, the municipal area is divided into eight planning and administrative regions (see Appendix F for population distribution per region from ). According to the overall crime rates, Region 8, which is the Inner City, has the highest total crime rate (> per persons). Three regions fall in the second highest category ( crimes per persons). They are Region 7 (Pretoria North), Region 3 (Eersterus/Mamelodi) and Region 4 (eastern Pretoria). Region 6 (Pretoria West) falls in the middle category (Map 16). The distribution of violent crime, however, follows a completely different pattern (Map 17). While the inner city also has the highest rate of violent crime (>600 per persons), Region 3 (Eersterus/Mamelodi) also falls within this category. None of the regions falls within the second category, while Region 6 (Pretoria West) falls within the middle category ( violent crimes per persons). Region 4 (Pretoria East) has the lowest violent crime rate (<150 per persons) in Tshwane. The distribution of property crime per region is exactly the same as that of the total crime rates per region in Tshwane, with the highest property crime rates in Region 8 (inner city) and the second highest in the surrounding regions 3, 4 and 7 (Map 18). The question arises whether there is any relationship between property crime and the distribution of enclosed neighbourhoods and security estates in Tshwane. According to the returned questionnaire, enclosed neighbourhoods occur mainly in Die Wilgers, Moreleta Park, Lynnwood Park, Wapadrand, Murrayfield, Faerie Glen, Lynnwood, Garsfontein, Waterkloof (all Region 4), Prinshof, Menlo Park, Nieuw Muckleneuk (all Region 8), Eldoraigne, Zwartkops, Irene, Wierda Park, Raslouw, Bronberrick (all Region 5). Region 8 has the highest crime rates in Tshwane and Region 4 falls in the category with the second-highest crime rate. Region 5, however, falls in the category with the second-lowest crime rate. Therefore, while crime may play a very significant role, especially in Regions 8 and 4, the prevalence of enclosed neighbourhoods in Region 5, as well, suggests that additional factors may affect the trend towards enclosing neighbourhoods in Tshwane. One of these may be the average income level within the regions, which will be investigated in more detail in the next section. According to the questionnaire, security estates were most prevalent in the Garsfontein and Wapadrand areas (both in Region 4), as well as in Highveld, Irene, Eldoraigne, Rooihuiskraal North, Kosmosdal (all Region 5). As mentioned before, Region 4 falls within the secondhighest category for property crimes, while Region 5 falls in the second-lowest category. This, again, suggests influencing factors other than high property crime rates. Apart from crime rates and average income, certain types of gated communities may be more likely in certain types of urban areas. While secure, high-rise apartment blocks may be more prevalent in inner city areas, enclosed neighbourhoods and security estates may be more prevalent in suburban areas. The response from the City of Tshwane indicated that enclosed neighbourhoods and security estates both occur mainly in the urban periphery. Tshwane indicated the existence of secure office parks, secure townhouse complexes and high-rise apartment blocks, as well. CSIR Building and Construction Technology 46

54 Map 16: Total crime rates per persons for the period in the City of Tshwane CSIR Building and Construction Technology 47

55 Map 17: Violent crime rates per persons for the period in the City of Tshwane CSIR Building and Construction Technology 48

56 Map 18: Property crime rates per persons for the period in the City of Tshwane CSIR Building and Construction Technology 49

57 4.2 Relationships with high-income areas The previous section suggests that there are significant factors in addition to crime, and more specifically property crime, that influence the occurrence of gated communities, in particular enclosed neighbourhoods and security estates. Since both of these types usually require large capital and operational resources, income may be a significant factor influencing the spread of these two types of gated communities in South Africa. This section will consider the relationship between enclosed neighbourhoods/security villages and the average income per capita per province, municipality and IDP region in the cities of Johannesburg and Tshwane Distribution of income and gated communities in provinces When comparing the income per capita (R), per province (Map 19), both Gauteng and the Western Cape fall into the category with the highest annual income per capita (R R23 455) (WEFA 2001) Five of the provinces fall within the middle category, namely Free State, Northern Cape, North West, Northern Province (now Limpopo) and Mpumalanga. KwaZulu-Natal and the Eastern Cape fall into the second-lowest category. Map 19: Average annual income per capita (2000) per province in South Africa If one compares the number of municipalities per province with enclosed neighbourhoods and security estates to the income per capita per province, Gauteng and the Western Cape have the highest income per capita and the highest number of municipalities (according to those municipalities that responded) with enclosed neighbourhoods and security estates. In other cases the relationship, however, appears to be weaker, as in the case of the Limpopo and KwaZulu-Natal. Limpopo has a relatively higher number of municipalities with both enclosed neighbourhoods and security estates, with an average income per capita, while KwaZulu-Natal falls in the second-lowest category in terms of income per capita, but it has CSIR Building and Construction Technology 50

58 the highest number of municipalities with security estates. As before, this suggests the presence of other factors influencing the distribution of these two types of gated communities per province, which may have to do with the size of urban conglomerations, crime levels, fear of crime, etc Distribution of income and gated communities in municipalities Map 20 indicates annual individual incomes 7 (in rand) in South Africa, by categorising all municipalities into five groups. Nine municipalities (3.7%) fall into the highest category. Six of these are located in Gauteng, two in the Western Cape and one in the Northern Cape. Gauteng and the Western Cape also recorded the highest income per capita in the country, as discussed in the previous section. Of these nine municipalities, eight (88.8%) responded to the questionnaire. Of those municipalities falling in the highest category for individual income, just over half (55.5%) indicated the existence of enclosed neighbourhoods, compared to two-thirds (66.6%) that had large security estates. In the case of enclosed neighbourhoods, 11.1% of these however indicated that they had received applications for additional enclosed neighbourhoods. One of the municipalities in this category did not respond. However, due to a small n factor for both enclosed neighbourhoods and security villages, as well as the fact that the map only indicates individual incomes per municipality, it is very difficult to establish for certain that there is a relationship between the occurrence of these two types of gated communities in South Africa and the level of individual income per municipality. Fifty-two (21.2%) of the municipalities fall into the second-highest category (Map 20), with the majority of these occurring in the Northern Cape (16), Western Cape (15), Northern Province and Gauteng (5). 7 It was not possible to use the average income per capita as in the case of the national map, since the data were severely influenced by what is generally considered the movement of people between municipalities to seek employment. CSIR Building and Construction Technology 51

59 Map 20: Total income per municipal area in South Africa CSIR Building and Construction Technology 52

60 4.3.3 Distribution of income and gated communities in City of Johannesburg and City of Tshwane The cities of Johannesburg and Tshwane were both in the highest category for annual individual income per municipality in South Africa. This is in addition to their very high population numbers and the highest crime rates (per metropolitan area) in South Africa. The previous discussion also suggested that there may be factors leading to higher numbers of enclosed neighbourhoods and security estates apart from crime rates in these municipalities. The larger IDP regions coincide with the ward boundaries, which constitute the units of measure for the income statistics. A number of wards therefore make up an IDP Planning and Administrative Region. The ward information was used to compile the maps to indicate the individual annual income per IDP region for both Johannesburg and Tshwane. a) Gated communities and income distribution in the City of Johannesburg Although Johannesburg falls within the category with the highest individual income per municipality, levels of individual income are not uniformly distributed throughout the municipal area. Only one IDP region falls within the highest category, with an individual annual income greater than R (Map 21). Three municipalities fall within the second category, namely Midrand/Ivory Park, Sandton/Rosebank and Alexandra. As mentioned before, the City of Johannesburg indicated that the majority of enclosed neighbourhoods and security estates occur in Randburg, Sandton, Alberton (in Johannesburg South), Midrand and Roodepoort (Question 9). Randburg is part of Roodepoort, which has the highest levels of individual income per IDP area in Johannesburg. Sandton and Midrand both fall in the second-highest categories. Alberton (as part of Johannesburg South), however, falls in the middle-income category under Johannesburg South. All the areas with a high prevalence of both enclosed neighbourhoods therefore fall in the top three categories for annual individual income. In the case of Alberton, one must also keep in mind that Alberton comprises only one part of the Johannesburg South area, which also includes lower-income areas. This may explain the low average levels of individual income. On the other hand, Randburg also includes mediumincome and lower-income areas, which means that Randburg s average individual income must be exceptionally high. Therefore, while these maps seems to indicate a relationship between income levels and the high prevalence of enclosed neighbourhoods and security villages in the City of Johannesburg, one will have to look at income levels on a smaller scale (possible ward level) as well, to confirm this pattern. CSIR Building and Construction Technology 53

61 Map 21: Annual individual income per IDP Region in the City of Johannesburg CSIR Building and Construction Technology 54

62 b) Gated communities and income distribution in the City of Tshwane As in the case of Johannesburg, income is not uniformly distributed in the City of Tshwane (Map 22). Two of the regions fall within the highest category for annual individual income (>27 000) namely Region 8 (Inner City, including old high-income areas), and Region 4 (Pretoria East). Only one area falls in the second-highest income category, namely Region 5 (Centurion). Region 7 (Pretoria North) falls in the third-highest category. The question arises whether there is any relationship between high levels of income and the distribution of enclosed neighbourhoods and security estates in Tshwane. As mentioned before, enclosed neighbourhoods mainly occur in Die Wilgers, Moreleta Park, Lynnwood Park, Wapadrand, Murryfield, Faerie Glen, Lynnwood, Garsfontein, Waterkloof (all Region 4), Prinshof, Menlo Park, Nieuw Muckleneuk (all Region 8), Eldoraigne, Zwartkops, Irene, Wierda Park, Raslouw, Bronberrick (all Region 5). This seems to indicate a strong relationship between higher income areas and the occurrence of enclosed neighbourhoods, since in the areas with a high prevalence of enclosed neighbourhoods there are three that fall within the two highest income categories. According to the questionnaire, security estates were most prevalent in the Garsfontein and Wapadrand areas (both in Region 4), as well as in Highveld, Irene, Eldoraigne, Rooihuiskraal North, Kosmosdal (all Region 5). As mentioned before, Region 4 falls within the highest category for annual individual incomes, while Region 5 falls in the second-highest category. This also seems to suggest a relationship between the high-income areas and the occurrence of large security estates. One explanation for the lower prevalence of large security estates in Region 8 is that these estates require large greenfield sites, which are mostly available only in peripheral suburbs (hence parts of Regions 4 and 5), while Region 8 mainly comprises older, mostly developed areas (including the CBD and well-developed older suburbs around it). 4.3 Relationships of type of gated community with the type of municipality or its location The previous two sections explored the distribution of enclosed neighbourhoods and large security estates in terms of their relationship with crime and higher-income areas, and the possibility of additional factors influencing their distribution. The question is whether the type or character of the municipality is a factor. The next section discusses possible relationships between different characteristics of municipalities and the prevalence of enclosed neighbourhoods and security estates. CSIR Building and Construction Technology 55

63 Map 22: Individual annual income per IDP region in the City of Johannesburg CSIR Building and Construction Technology 56

64 4.3.1 Urban/non-urban municipalities The terms urban and rural can be confusing or contested, and therefore the terms urban and non-urban are often used, as they are by Stats SA and other leading population databases such as PIMSS ( In terms of urban population per province, Gauteng (97%) is by far the most urbanised (see Figure 4.1), followed by the Western Cape (89%). At the other end of the spectrum is Limpopo Province (formally known as Northern Province), with an urban population of only 11% Nonurban Urban 0 Gauteng Western Cape Northern Cape Free State KwaZulu-Natal Mpumalanga Eastern cape North West Northern Province South Africa Figure 4.7: Percentage urban population per province When analysing the proportion of urban population in the 237 local and metropolitan municipalities in South Africa, in terms of whether they comprise more or less than 50%, the outcome is very evenly distributed (Figure 4.2). 50% and higher urban population 51% < 50 % urban population 49% Figure 4.8: Distribution of municipalities in South Africa with more or less than a 50% urban population CSIR Building and Construction Technology 57

65 If one however makes use of five categories, the numbers are disproportionate (Figure 4.3). The largest proportion (33%) of all municipalities have an urban population of under 20% and are thus predominantly non-urban, including farm land and farmsteads, scattered settlements, rural villages and smaller rural towns. The second-highest (25%) category is at the other extreme of the scale, namely municipalities with an urban population of between 81 and 100%, followed closely (20%) by those municipalities with urban populations of between 61 and 80% % 25% 0-20 % 33% Figure 4.9: Distribution of % urban population in municipalities in South Africa 61-80% 20% % 11% % 11% The question is whether the occurrence of enclosed neighbourhoods and security estates is influenced by the proportion of urban population in a municipal area, since enclosed communities are generally regarded as predominantly an urban phenomenon. The highest proportion of enclosed neighbourhoods (52%) reported by local authorities occurred in the municipalities with the greatest urban populations (see Table 4.3). This means that more than half of all the municipalities indicating the existence of enclosed neighbourhoods have more than an 80% urban population. Enclosed neighbourhoods were, however, not restricted to this category. A few municipalities with an urban population of less than 20% also indicated the existence of enclosed neighbourhoods. Here other factors may play a role, such as the presence of a large city within that municipal area. Table 4.3: Distribution of enclosed neighbourhoods and security estates per category of percentage urban population in municipalities in South Africa % Urban population in municipality Enclosed neighbourhoods Security estates N % N % N % 0-20 % % % % % Missing A similar pattern was found in the case of security estates. The highest proportion (43%) of municipalities with security estates was also found in the group with the greatest urban populations. Correlations, however, showed no significant correlation between the CSIR Building and Construction Technology 58

66 percentage of urban population and the municipalities with enclosed neighbourhoods, or between urban population and the number of road closures per municipality (see Appendix H). There was, however, a positive relationship and significant correlation between urban population and the number of security estates (0.0063). One cannot, however, consider proportion alone, since the size of urban areas is also likely to influence the distribution of gated communities, due to the concentrations of people Sizes of urban areas in municipalities The proportion of urban population determines not only the nature of a municipality, but also the size of cities or towns within municipal areas. For example, one municipal area can have 10 small towns scattered over the geographical area, with people in each, contributing to an urban population of in the municipal area. Another can have only two urban areas, one small town of people and one larger town of , also with a total of people in the municipal area. In both cases the total urban population would therefore be the same. Larger concentrations of people and higher densities could give rise to different socio-economic conditions, for example levels of crime. Worldwide, crime rates have been shown to be higher in cities than in rural areas, with the rate usually increasing according to city size, because most factors associated with high crime rates are more prevalent in cities than in small towns (Louw and Schönteich 2001). The old transitional municipalities in South Africa were divided into urban municipalities (TLCs 8 ) and rural municipalities (TRCs 9 ), but the new municipalities encompass both urban and rural areas in an attempt to integrate urban and rural development. One of the consequences of this demarcation is that whereas in the past TLCs included only urban areas, and therefore meant that one could use these to compare the sizes of urban areas or different cities and towns, this is not possible with the new municipal areas. One of the ways to overcome this problem was to plot all the urban areas in South Africa and then divide the municipalities in groups, according to the largest cities or towns in their geographical area. This would then indicate municipalities with large cities, medium cities, large towns, small towns, etc (see Map 23). The sizes of urban areas are indicated by different sizes of dots, and the different groups of municipalities are graded (the darkest blue having the largest cities in their areas). The six metropolitan municipalities are also indicated as those with a population smaller than 1.5 million and those with a population larger than 1.5 million. Gauteng has a number municipalities with large cities or urban areas, making it the most urbanised province (97%) (see Figure 4.7). 8 Transitional local councils 9 Transitional rural councils CSIR Building and Construction Technology 59

67 Map 23: Sizes of urban areas within municipal areas CSIR Building and Construction Technology 60

68 Comparing the spatial distribution of the enclosed neighbourhoods with the location of larger cities and municipalities that have larger urban populations indicates that there is some form of relationship. Out of the six large urban conglomerations (metropolitan municipalities) in South Africa, each with a population of over one million people, all have road closures and four indicated high numbers of road closures. The highest numbers were recorded in the three metropolitan municipalities in Gauteng, two of which have a population greater than 1.5 million people. A growing number of illegal road closures are also occurring in e Thekweni. However, this pattern does not always seem to be the case in municipalities with smaller cities/towns. Although enclosed neighbourhoods tend to occur in larger cities/towns, there are some exceptions. Given the larger urban populations in Bloemfontein, Welkom and East London, one would expect a number of road closures in these cities, whereas there are currently none. One explanation is that the Mangaung Municipality (including Bloemfontein) does not allow road closures. According to local officials, the demand for them is, however, growing. Other areas with larger urban conglomerations are more rural-based and can be classified as dispersed urban settlements, which often include large concentrations of people that were located in former homelands areas, such as Bushbuckridge, Thohoyando, etc. On the other hand, there are a few municipalities with towns of less than people that have reported the existence of road closures. These include the towns of Colesberg, Nylstroom, Naboomspruit, Musina, Giyani, Groblersdal, Marble Hall, Mooi River, Greytown, Melmoth, etc. This suggests that other aspects probably influence the distribution and growth of enclosed neighbourhoods there as well, such as crime, fear of crime, etc. A significant correlation (with a probability value of ) was shown between the size of urban areas and municipalities with existing road/neighbourhood closures (see Appendix I for tables). This indicates a strong positive relationship between the size of the city/town in a municipal area and the existence of enclosed neighbourhoods. There also appears to be some form of relationship between security estates and large urban areas, when one compares the spatial distribution of these two variables. Out of the six metropolitan areas, four indicated the existence of large security estates and three reported very large numbers, namely the cities of Tshwane, Johannesburg and Cape Town. Other larger cities/towns with higher numbers of security estates (when the different numbers of security estates within municipalities that responded are compared) include Vereeniging and Vanderbiljpark (> people each). This shows that a large proportion of metropolitan municipalities have recorded the existence of both types of development. One of the medium cities/towns with a large numbers of security estates is that of Knysna. As with enclosed neighbourhoods, there are also a few municipalities with smaller towns that have reported high numbers of security estates, such as Plettenberg Bay, Hartbeespoort and Lydenburg, all with populations of less than people. This suggests that there are probably other factors influencing the distribution and growth of enclosed neighbourhoods, such as living in a secure environment of natural beauty. The probability of strong additional predisposing factors is also supported by the fact that no significant correlation was found between the number of security estates and the sizes of urban areas Metropolitan and non-metropolitan municipalities As stated above, municipalities in South Africa can be categorised as district, metropolitan and local municipalities. District municipalities comprise a number of local and/or metropolitan municipalities and, as mentioned before, were not included as a unit of measure in this study. Local and metropolitan municipalities were. The question is therefore whether there is a correlation between enclosed neighbourhoods and security estates and CSIR Building and Construction Technology 61

69 metropolitan/non-metropolitan municipalities. One would assume that the highest numbers of both these types occur in metropolitan areas. Yet a significant negative correlation was indicated between the number of enclosed neighbourhoods and metropolitan municipalities, as well as between the number of security estates and metropolitan municipalities (both with p-values = <.0001). This may be explained by the proportion of metropolitan municipalities relative to non-metropolitan municipalities. Metropolitan councils make up only 2.5% of the total number (237) of municipalities under consideration (metropolitan and local municipalities in South Africa). If one compares the number of municipalities (metro/non-metro) with existing enclosed neighbourhoods and security estates, Table 4.4 shows that the 17% which indicated the existence of enclosed neighbourhoods were metropolitan municipalities, compared to 14% of those with security estates. Both of these therefore had higher percentages of yes responses than the percentage metropolitan municipalities (2.5%). However, if one compares the percentage of metropolitan municipalities having enclosed neighbourhoods and security estates with the percentage of non-metros, 66% of the metropolitan areas have enclosed neighbourhoods compared to 8% of the non-metro areas. The same pattern is evident in terms of security estates: 66% of metros have security estates, compared to 10% of the non-metros. Table 4.4: Metro/non-metro municipalities with enclosed neighbourhoods and security estates Municipalities Enclosed neighbourhoods Security estates No % No "yes" % of total % of group No "yes" % of total % of group Metro Local Total Note: e'thikweni (metro) has both security estates and illegal closures 10 However, as before, there were a few discrepancies in terms of the replies to different questions. If one compares the responses from six metropolitan municipalities, this becomes apparent (Table 4.5). Four of the respondents indicated the existence of enclosed neighbourhoods and five indicated the existence of security estates. The City of Johannesburg and the City of Tshwane both indicated the existence of both types and reflected the highest numbers of road closures. The City of Cape Town, however, had higher numbers of security estates (see Appendix E for map of security estates in Cape Town). Ekurhuleni had high numbers of road closures. Nelson Mandela responded that it did not have existing neighbourhood closures. However, in a previous survey, it indicated the existence of neighbourhood closures, but based on the private approach to development. e Thekwnini also indicated that no enclosed neighbourhoods existed in that area. Other research, however, has confirmed that, while it does not allow the legal closure of roads for security purposes, it has many illegal road closures in the metropolitan area, especially in the western areas around Westville. 10 The data for e Thekweni were skewed by the fact that the questionnaire was distributed to all the sub-structures of the metropolitan area, and not all of them responded. It therefore does not represent a complete picture of the situation. Additional data were obtained through follow-up phone calls. CSIR Building and Construction Technology 62

70 Table 4.5: Metropolitan municipalities with enclosed neighbourhoods and security estates Metros Responded "yes" Actual numbers Encl neigh Sec est RC Illegal SE Tshwane Yes Yes 36 Yes 18 Johannesburg Yes Yes 300 Yes 20 Nelson Mandela No Yes 0 No Missing e'thekwini No Yes 0 Yes Missing Cape Town Yes Yes 18 Yes 24 Ekurhuleni Yes No 25 No 2 RC = Road closures SE = Security estates Coastal / inland municipalities A first glance at the spatial distribution (Map 4) of security villages reveals a possibility that there may be a strong relationship between the number of security estates and coastal municipalities. The same map (Map 2) for enclosed neighbourhoods however, does not suggest a similar pattern. When comparing the distribution of security estates and the type of municipality in terms of coastal or inland, it is important to keep in mind that coastal municipalities only make up 10% of the total number of municipalities in South Africa. 10% Coastal Inland Figure 4.10: Comparison between percentage coastal and inland municipalities 90% If one compares the number of municipalities with existing enclosed neighbourhoods and security estates in coastal and inland municipalities (Table 4.4), 21.7% of the municipalities with enclosed neighbourhoods were coastal, compared to 28,6% of the municipalities with security estates. Given that only 10% of the municipalities are coastal, it is interesting that 28.6% of those that responded positively in terms of security estates were from coastal towns. If one, however, considers the percentage of coastal municipalities with enclosed neighbourhoods (20%), this is higher than the percentage of inland municipalities with enclosed neighbourhoods (8.45%). The same is true of security estates, where 33.3% of the coastal municipalities reported the existence of security estates, compared to 9% of inland municipalities. It is also interesting that the percentage of coastal municipalities with security estates (33%) is higher than those with enclosed neighbourhoods (20%). Table 4.6: Coastal/Inland municipalities with enclosed neighbourhoods and security estates Municipalities Enclosed neighbourhoods Security estates No % No "yes" % of total % of group No "yes" % of total % of group Coastal Inland Total CSIR Building and Construction Technology 63

71 If one compares the number of municipalities with one or more security estates 11, using the same categories as on Map 2, Table 4.5 indicates that 32% of those municipalities that responded were from coastal municipalities, while only 10% of the total number of municipalities are coastal. The municipalities with the highest numbers of security estates were evenly distributed between coastal and inland municipalities. Table 4.5: Distribution of number of security estates in coastal/ inland municipalities Coastal/inland municipalities Number of municipalities with number of security estates Type No % SE SE SE SE >20 SE Total % Coastal Inland Missing The correlation between the actual numbers of security estates and coastal municipalities is however not strong. This may be because the distribution is not evenly spread and the coastal municipalities only constitute 10% of the total number of municipalities in South Africa Municipalities with popular tourist destinations / natural features The previous section indicated that the highest number of security villages is evenly spread between coastal and inland municipalities. This raises the question whether this indicates the existence of another significant variable influencing the proliferation of security estates. Such a variable may be the presence of significant tourist attractions or natural amenities. Four of the six municipalities with very high numbers of security estates have either or both of these features. The municipalities of Plettenburg Bay and Knysna are located in areas greatly renowned for their natural beauty. Plettenberg Bay and Knysna are also located next to the coast. The growth of security estates may be related to the number of people retiring at the coast or having a second holiday home at the coast. In both of these categories, the choice of a home inside a security estate would appear to reduce the vulnerability of elderly people and those far away from their properties for long periods of time (second-home owners). The other two municipalities with very high numbers of security estates are located in Gauteng and the North West Province, namely Emfuleni and Madibeng. The Vaal River makes up the southern border of Emfuleni, which may result in an increased number of security estates next to the river. In the case of Madibeng, the majority of the security estates are clustered around Hartbeespoort Dam. One explanation may be the proximity of the dam to Pretoria and Johannesburg. Hartbeespoort Dam is a major attraction and recreational amenity, making it a popular location for many people working in Johannesburg or Pretoria (daily commuters) or those having weekend houses there. This chapter has indicated likely relationships between crime, fear of crime, high income, sizes of urban areas and, to a lesser degree, metropolitan areas, coastal towns (in the case of security estates) and gated communities in South Africa. The next chapter will discuss 11 The total number of such municipalities is 34, compared to the 28 that responded that they had existing security estates. This is due to the fact the six municipalities responded that they had one or more security estates, but did not indicate a yes to the earlier question, as explained in Chapter 3. CSIR Building and Construction Technology 64

72 government s response to gated communities in South Africa and more specifically enclosed neighbourhoods. CSIR Building and Construction Technology 65

73 CHAPTER 5: GOVERNMENT RESPONSE TO GATED COMMUNITIES 5.1 Policies on types of gated communities As mentioned in Chapter 2, security estates are private developments, while enclosed neighbourhoods are concerned with restriction of access to existing public roads. The legal and policy implications of these two types will therefore be different. While large security estates need large greenfield or brownfield sites, and therefore zoning permission to be developed (usually by a large developer), residents from enclosed neighbourhoods need to obtain permission from the local municipality to restrict/prohibit access into existing neighbourhoods. As discussed previously, this can take the form of a public approach, private approach or combination of the two. Currently there is no national policy on enclosed neighbourhoods, or road closures for security purposes as they are more generally known, in South Africa. 1 Planning and development in the country are guided by the White Paper on Spatial Planning and Land use Management (2001) and the Development Facilitation Act (1995), which will be replaced by the Land Use Act in the near future. Neither of these mention, or make any reference to, gated communities, road closures or neighbourhood enclosures, or security estates. There is also a whole range of other national policies and legislation that has an impact on planning and development to a greater or lesser degree. Only one province makes provision for road closures for security purposes at a provincial level. The provincial Government of Gauteng incorporated a chapter (Chapter 7) in the Rationalisation of Local Government Affairs Act, No. 10 of 1998, to deal with access restrictions for security purposes. The Act came into operation on 19 March Chapter 7 makes provision for a local authority to restrict access into an area for purposes of enhancing safety and security. 2 The local authority may restrict access on its own initiative, or may authorise any legal body or institution to do so, as in the case of neighbourhood closures. The existence of this Act does not, however, solve all the legal questions regarding the application process for road or neighbourhood enclosures, or technical details of what is required or allowed. This is up to the local municipality to define and describe. When one considers the existence of policies at a local government level, the picture changes somewhat. The survey indicated that 14 of the respondents (12%) had policies on road/neighbourhood closures (Figure 5.1), compared to 23 (19%) that merely reported the existence of enclosed neighbourhoods. Therefore, not all municipalities with enclosed neighbourhoods have policies in place to guide their approval and establishment. Previous research has indicated that one of the reasons for denying applications for road/neighbourhood closures was often that the relevant local municipality did not have a policy in place (Landman 2000). 1 Since the term gated communities is not yet well established in South Africa and since there is not always agreement on what the term encompasses (both in South Africa and internationally), few government agencies consider security developments as a collective at this stage. Consequently there are almost no policies on gated communities as such or security developments, although a number of local SDFs refer to the growth of these types of development in the spatial analysis of their local areas. 2 The Department of Local Government in Gauteng is in the process of completing a new policy on enclosed neighbourhoods. CSIR Building and Construction Technology 66

74 The majority of these policies were developed in 1998 (36%) and 2000 (29%) (Figure 5.2), thus fairly recently. One of the respondents replied that their policy had been developed in It is not clear whether it was a policy on the closure of roads for security or normal transport purposes. The latter may explain the development of a policy at such an early date. N.A. 30% Missing 1% Yes 12% No 57% % Missing 14% % % % Figure 5.1: Policies on road/neighbourhood enclosures Figure 5.2: Year that the policies were developed A number of respondents (15%) indicated that they were in the process of developing or revising their policies on road/neighbourhood closures (Figure 5.3). Some of these municipalities already had a policy in place and were revising it (17.65% of those with policies), while others were developing new policies (19.40% of those preparing/revising policies). N.A. 43% Missing 1% Yes 15% Figure 5.3: Municipalities in the process of revising or developing policies on road/neighbourhood closures Don't know 1% No 40% Previous research showed that road/neighbourhood closures and security estates can have a significant impact on spatial transformation, as well as urban planning and development (Landman 2000; Landman 2000b; Landman 2002). The questionnaire therefore also included questions on the consideration of controlled access areas in broader urban planning policies as well as the IDPs (Questions 6 to 8). Fifteen of the respondents (13%) indicated that they considered any form of enclosed area with controlled access a major issue in their local urban policies. What is interesting is that only seven of these indicated that they had existing enclosed neighbourhoods in their municipal areas. On the other hand, 15 of those municipalities that indicated the existence of enclosed neighbourhoods in their municipal areas did not consider any form of enclosed area with controlled access a major issue in their local urban policies. All local municipalities in South Africa are required to develop an Integrated Development Plan (IDP) to inform planning and development decision-making in their municipal areas. CSIR Building and Construction Technology 67

75 This requires an in-depth analysis of all aspects influencing planning and development, the development of a number of development goals, and prioritising strategic actions for immediate intervention. The question is whether any form of enclosed area with controlled access has been addressed as a major issue in the IDPs. Only five (4.27%) municipalities among those that responded indicated that they had considered any form of enclosed area with controlled access a major issue in their local IDPs. Of these, only 2 (40%) indicated the existence of enclosed neighbourhoods in their municipal areas. On the other hand, 21 of those that indicated the existence of enclosed neighbourhoods in their areas (91.35%) did not consider any form of enclosed area with controlled access a major issue in their local IDPs. From this it appears that very few municipalities have considered the issue of controlled access a major issue in their broader urban development policies or plans at this point, and that only a few have policies in place to guide the development of such areas. It may, however, be that the growing demand will force a change in this regard in the future. 5.2 Relationships between policy development, applications and approvals for types of gated communities and types/location of municipalities The survey indicated that 38 of the respondents (32%) had received applications for road/neighbourhood enclosures (Figure 5.4). Half of these had approved one or more of the applications (Figure 5.5). Yes 32% Missing 3% No 68% Denied 47% Approved 50% Figure 5.4: Percentage of respondents with applications for road/neighbourhood closures Figure 5.5: Percentage of approvals of respondents with applications for road/neighbourhood closures Of those municipalities that had received applications for road/neighbourhood enclosures, 44.74% had received only one application and 23.68% two applications (Table 5.1). More than two-thirds (68.28%) of the municipalities that had received applications had therefore received fewer than three applications. Six municipalities had received more than five applications and three more than 10. These three had received very high numbers of applications namely 24, 79 and 81. Table 5.1: Number of applications received per municipality No of applications Total No of municipalities % of total CSIR Building and Construction Technology 68

76 Of the six municipalities that had received more than five applications, three were metropolitan municipalities and five located in Gauteng. The three having received more than 10 applications were the three metropolitan municipalities in Gauteng: Ekurhuleni (24), Johannesburg 3 (79) and Tshwane (81). It therefore appears that there is a high demand for neighbourhood enclosures in Gauteng. The may relate to high total crime rates, property crime rates and income per capita in the province. If one compares the number of applications with the number of approvals, it becomes evident that not all applications are approved. While some municipalities may approve all applications, others may approve none, and others approve only some of those meeting their requirements. For example, the responses showed that, out of the 16 municipalities (43%) that had received one application for a neighbourhood closure, only six had approved it. Out of the nine municipalities that had received two applications each, six had approved none, two had approved one, and one had approved two. At the other end of the scale, one municipality had received 79 applications, approving 23, while another had received 81 applications, approving only 47. High numbers of applications do therefore not necessarily result in high numbers of approvals, although this appears to occur in most cases (Figure 5.6). There is a strong relationship between the number of approvals and the number of applications, apart from one outlier to the left of the graph. This represents the City of Tshwane, which has suddenly received a very high number of applications, giving rise to a considerable backlog in terms of approvals. This is reflected in Figure 5.6, where Tshwane constitutes the outlier. Apart from Tshwane, there is a strong relationship between the number of applications and the number of approvals within the relevant local municipalities Number of applications Number of approvals Figure 5.6: Relationship between the number of applications and number of approvals for road/neighbourhood enclosures 3 The number of applications (79) refers only to those received after the establishment of the City of Johannesburg in December 2000, and not to those received by the previous municipality responsible for Johannesburg, namely the Greater Johannesburg Metropolitan Council, which excluded Midrand. CSIR Building and Construction Technology 69

77 This, however, does not seem to be the case in the City of Johannesburg. There appears to be a strong relationship between the number of approvals and the number of applications (top right corner). There may be two reasons for this. First, the majority of the applications were received and processed by the previous municipality (GJMC). Secondly, Johannesburg has been involved in a lengthy process to reconsider its policy on road closures since the establishment of the new municipality in December Since then, there has been a moratorium on the approval of road closures for safety in the municipal area. One of the major consequences of this has been that most interested parties have not bothered to apply. According to an interview with an official, only about 5 to 10% of all the enclosed neighbourhoods that have since been established have formally applied for permission from the council. The rest have resorted to illegal closures. Only five (4.27%) of the municipalities that responded reported the existence of illegal road closures in their municipal areas. However, it is estimated that there are currently 188 illegal road closures in the City of Johannesburg (Access Restriction Policy of the City of Johannesburg, 2003). 5.3 Policies on gated communities in the cities of Johannesburg and Tshwane The City of Johannesburg adopted a new policy in April The result was a three-month period of grace (till the end of July 2003) for all illegal closures to apply. After this period all illegal closures were to be removed. The City of Johannesburg constitutes the former GJMC (including four local transitional councils) and the Midrand Local Council, which used to be part of the Khayalami Metropolitan Council. Midrand and two of the previous substructures, the Eastern Metropolitan Local Council (EMLC) and the Northern Metropolitan Local Council used to have policies in place. The City of Johannesburg now functions as a unicity and has therefore needed to have one policy in place for the entire area. Due to a split in the council, as well as several objections to road closures, the process of developing a new policy took almost two years. The process of developing a new policy also included a series of public hearings on security access restriction held in August 2002, as well as a number of workshops with important role-players. The City of Tshwane currently does not have an approved policy in place. As with the case in Johannesburg, two of the previous three local councils that constituted the former Greater Pretoria Metropolitan Council had policies in place, namely the Pretoria Local Council and the Centurion Local Council. After the reconstruction and the amalgamation of a number of local councils into the City of Tshwane, the council embarked on developing a policy, which was completed towards the end of However, due to several objections and some problems with the 2002 policy, Tshwane embarked on a process to draft a second, more upto-date policy. It also embarked on a series of public hearings to assist the process and invite comments on the 2002 policy at the end of March The revised policy is currently being formulated and has not yet been approved by the Council. Therefore, the moratorium on the approval of any applications in the City of Tshwane is still in place. CSIR Building and Construction Technology 70

78 CHAPTER 6: CONCLUSION AND WAY FORWARD Gated communities have rapidly increased in the past few years. This has happened in a context of great municipal transformation (institutional and spatial), as well as very high crime rates. This study aimed to determine the location and extent of gated communities in the country, possible influences on the distribution and the government s response to these types of developments, in order to identify lessons for decision-makers and policy implications. This was done on various scales. The national distribution was analysed in terms of provinces and municipalities. It also included a focus on the situations in the cities of Johannesburg and Tshwane in Gauteng. Although the study highlighted a number of useful and relevant findings at all these scales, it seems that the findings within municipal areas were more convincing than those between municipalities and provinces. 6.1 The location and extent of gated communities in South Africa Although the total proportions of local municipalities with enclosed neighbourhoods (20%) and security estates (23%) are more or less the same (according to those that responded), the distribution of these two types per province differs substantially. Gauteng and the Limpopo Province had the highest number of local municipalities with enclosed neighbourhoods, followed closely by the Western Cape. On the other hand, KwaZulu-Natal and the Western Cape had the highest number of local authorities with security estates, followed by Gauteng and Limpopo Province. It must however be kept in mind that the number of local municipalities per province differs substantially, and this therefore needs to be corrected in some way to be able to make true comparisons. Three provinces, Gauteng, Limpopo Province and Western Cape had higher numbers of local authorities with both enclosed neighbourhoods and security estates. Gauteng had the highest number of municipalities with enclosed neighbourhoods and security estates. The numbers of enclosed neighbourhoods and security villages within municipalities differ substantially. This is true both for the number of one type in different municipalities, as well as the distribution of both types in municipal areas. The survey indicated that the highest numbers of enclosed neighbourhoods occurred in Gauteng. Two of the three metropolitan municipalities had the highest numbers of road closures at the time of the response, namely Tshwane and Johannesburg. There were also two municipalities in the Western Cape with higher numbers, namely those including Cape Town and Mossel Bay. There are also a number of illegal closures in many municipalities. Five municipalities were aware of illegal neighbourhood closures in their areas. There were reported to be about 200 illegal road closures in the City of Johannesburg at the time. The distribution of security estates differs to some extent. While some of the municipalities mentioned also had higher numbers of security estates, such as the cities of Johannesburg, Tshwane and Cape Town, higher numbers are occurring in other municipalities that recorded relatively low numbers of road closures, or none at all. The highest numbers of security estates were recorded in the Emfuleni and Madibeng municipalities, in Gauteng and the North West province. Other municipalities with higher numbers of security estates include the Plettenberg Bay and Knysna municipalities. They are both in the Western Cape. According to the questionnaire, the highest numbers per municipality, within cities, for both enclosed neighbourhoods and large security estates, occurred in peripheral suburbs, compared to inner-city areas, inner suburbs and traditional townships. CSIR Building and Construction Technology 71

79 6.2 Explaining the distribution If one compares the spatial distribution of the enclosed neighbourhoods with the location of larger cities and municipalities that have larger urban populations, it appears that there is some form of relationship. Out of the six large urban conglomerations (metropolitan municipalities) in South Africa, all with a population of over one million people, all have road closures and four indicated the existence of higher numbers of road closures. The highest numbers were recorded in the three metropolitan municipalities in Gauteng, two of which have a population greater than 1.5 million people. A growing number of illegal road closures are also occurring in e Thekweni (including Durban). However, this pattern does not always seem to be the case in municipalities with smaller cities/towns. Although enclosed neighbourhoods tend to occur in larger cities/towns, there are some exceptions. Given the larger urban populations in Bloemfontein, Welkom and East London, one would have expected a number of road closures in these cities, where none were reported. One explanation is that the Mangaung Municipality (including Bloemfontein) does not allow road closures. According to discussions with local officials, the demand for them is, however, growing. Other areas with larger urban conglomerations are more rural-based and can be classified as dispersed urban settlements, which often include large concentrations of people that were located in former homelands areas, such as Bushbuckridge, Thohoyando, etc. On the other hand, there are a few municipalities with towns of fewer than people that have reported the existence of road closures. These include the towns of Colesberg, Nylstroom, Naboomspruit, Musina, Giyani, Groblersdal, Marble Hall, Mooi River, Greytown and Melmoth. This suggests that other aspects influence the distribution and growth of enclosed neighbourhoods here, such as crime and the fear of crime. There also seems to be a relationship between security estates and large urban areas, when one compares the spatial distribution of these two variables. Out of the six metropolitan areas, four indicated the existence of large security estates and three reported very large numbers, namely the cities of Tshwane, Johannesburg and Cape Town. Other larger cities/towns with higher numbers of security estates include Vereeniging and Vanderbiljpark (> people each). Medium cities/towns with larger numbers of security estates include Knysna. As with enclosed neighbourhoods, there are also a few municipalities with smaller towns that have reported high numbers of security estates, such as Plettenberg Bay, Hartbeespoort and Lydenburg, all with a population of less than people. This suggests that other aspects influence the distribution and growth of enclosed neighbourhoods here as well, such as living in a secure place amid the beauty of nature. The probability of strong additional influences is supported by the fact that no positive relationship was found between the number of security estates and the sizes of urban areas. Another strong influencing factor is the crime rate and more specifically property crime. As mentioned earlier, property crimes are often related to high-income areas, which in addition have the resources to respond to crime through more physical protection, such as fencing off areas or choosing to stay in walled villages. According to the survey, 85% of those municipalities that fell into the highest category for property crimes (>550 per population) had either enclosed neighbourhoods or security estates, while 40% had both. This seems to indicate a relatively strong relationship between high crime rates and the existence of enclosed neighbourhoods and security estates as a way to respond to these crimes. This study has indicated relationships between crime, high income, sizes of urban areas and to a lesser degree metropolitan areas, coastal towns (in the case of security estates) and gated communities in South Africa. These issues are influencing the distribution and growth of gated communities in the country and will probably continue to do so in the future. CSIR Building and Construction Technology 72

80 6.3 The government s response to gated communities Currently there is no national policy on enclosed neighbourhoods, or road closures for security purposes, as they are more generally known in South Africa. Planning and development in the country are guided by the White Paper on Spatial Planning and Land use Management (2001) and the Development Facilitation Act (1995), which will be replaced by the Land Use Act in the near future. Neither of these mention, or make any reference to, gated communities, road closures or neighbourhood enclosures, or security estates. This leaves the path open for provinces and local councils to respond in their own way. Only one province makes provision for road closures for security purposes at a provincial level. The provincial Government of Gauteng incorporated a chapter (Chapter 7) in the Rationalisation of Local Government Affairs Act (No 10 of 1998) to deal with access restrictions for security purposes. The Act came into operation on 19 March 1999 and is still guiding the formulation of local policies in this regard. It is, however, being reconsidered by the Department of Provincial and Local Government at the moment, and may be replaced by an updated policy on the restriction of access in the New Planning Bill that is to be released in Gauteng in the near future. As the demand for road closures grows in other provinces as well, they will also have to consider provincial guidance. When considering the existence of policies at a local government level, the picture changes somewhat. The survey indicated that 14 of the respondents (12%) had policies on road/neighbourhood closures, compared to 23 (19%) that reported the existence of enclosed neighbourhoods. Therefore, not all municipalities with enclosed neighbourhoods have policies in place to guide their approval and establishment. Previous research has indicated that one of the reasons for denying applications for road/neighbourhood closures was often that the local municipality did not have a policy in place. The majority of these policies were developed in 1998 (36%) and 2000 (29%)thus fairly recently. A number of respondents (15%) indicated that they were in the process of developing or revising their policies on road/neighbourhood closures (Figure 5.3). Some of these municipalities already had a policy in place and were therefore revising it (17.65% of those with policies), while others were developing new policies (19.40% of those preparing/revising policies). Another issue of importance was to what extent municipalities had considered any form of gated communities as part of their larger urban policies or IDPs. Fifteen of the respondents (13%) indicated that they considered any form of enclosed area with controlled access a major issue in their local urban policies. What is interesting is that only seven of these (46.67%) indicated that they had existing enclosed neighbourhoods in their municipal areas. On the other hand, 15 of those municipalities that indicated the existence of enclosed neighbourhoods in there municipal areas (68.18%) did not consider any form of enclosed area with controlled access a major issue in their local urban policies. Only five (4.27%) municipalities of those that responded indicated that they had considered any form of enclosed area with controlled access a major issue in their local IDPs. Of these, only two (40%) indicated the existence of enclosed neighbourhoods in their municipal areas. On the other hand, 21 of those that indicated the existence of enclosed neighbourhoods in their areas (91.35) did not consider any form of enclosed area with controlled access a major issue in their local IDPs. From this it appears that very few municipalities consider the issue of controlled access a major issue in their broader urban development policies or plans at this point in time, and only a few have policies in place to guide the development of such areas. It may however be that the growing demand will force a change in the future. If one compares the number of applications with the number of approvals, it becomes evident that not all applications are approved. While some municipalities may approve all applications, others may approve none, and others approve only some of those meeting the CSIR Building and Construction Technology 73

81 requirements. The survey indicated that 38 of the respondents (32%) had received applications for road/neighbourhood enclosures (Figure 5.4). Half of these had approved one or more of the applications (Figure 5.5). Of those municipalities that received applications for road/neighbourhood enclosures, 44.74% received only one application and 23.68% two applications. More than two-thirds (68.28%) of the municipalities that received applications therefore had received less than three applications. Six municipalities received more than five applications and three more than 10. These three received very high numbers of applications for neighbourhood closures namely 24, 79 and 81. High demand does not necessarily lead to rapid response. While the highest number of existing neighbourhood closures occur in the cities of Johannesburg and Tshwane, as well as the highest number of applications for neighbourhood enclosures, it took a long time (almost two years) for both councils to complete policies in this regard. This is partly due to the fact that it is a very contentious issue and there are strong protagonists and antagonists, making it difficult for local governments to respond to the needs of all urban residents. This section indicates that the issue of gated communities will have different implications for the three spheres of government in South Africa, but will at the same time require cooperation between them to ensure an integrated and coordinated approach. 6.4 Key findings and policy implications This study has lead to a number of key findings for those having to deal with both enclosed neighbourhoods and security villages in South Africa, including the following: The distribution of different types of gated communities differs between municipalities. For example, a large number of enclosed neighbourhoods does not necessarily mean a large occurrence of security estates, although this may occur in some municipalities. There are also differences between the types of gated communities in terms of their location within cities. There are different influences on different types of gated communities. The relevance and importance of the issue of gated communities to local councils differs. Local councils do not respond in similar ways to enclosed neighbourhoods There has been limited consideration of the issue in larger urban planning and development policies at a local government level The study indicated that the two types of gated communities which were studied in more depth, namely enclosed neighbourhoods and security estates, are not uniformly distributed throughout the country. This is also true of their extent, which differs substantially. This is likely to have a major impact on local councils response and importance given to the issue. It has also been shown that these two types are different in nature and not always influenced by the same factors. Where enclosed neighbourhoods are often influenced by high levels of property crime and areas with a high average income, security estates are perhaps less dependent on the actual crime rate, but more on a perception of crime in general, combined with a search for areas with beautiful natural amenities (such as the sea, dams, mountains, or rivers) or man-made attractions (including golf courses, etc). Since they are also established in different ways, this could also play a role such as the tendency of enclosed neighbourhoods to occur in existing well-established upmarket areas (brownfield sites) and security estates on the urban periphery where large greenfield sites are available. CSIR Building and Construction Technology 74

82 The study also showed that different local councils responded in different ways and that the issue of gated communities (even in areas with a comparatively high demand) does not always receive immediate attention and/or action. This is often due to a lack of consensus among the councils. The fact that very few local authorities consider these types of developments in their larger urban development policies often contributes to a lack of consistency regarding urban development at a local level. These aspects have a number of policy implications. Local councils cannot deal with applications for either enclosed neighbourhoods or large security estates if they do not have any policy in place or are unwilling to take a position. This could entail either developing a policy or deciding to refuse it on the basis of existing planning policies. It is very difficult to take a strong position on a specific development issue such as gated communities if it is not considered within the context of the larger urban development frameworks and policies. This is especially true if issues only concern certain parts of the metropolitan area. It is necessary to consider the impact on the city (metro) as a whole as well. A lack of consensus at provincial level or a lack of proper guidance can often lead to a very haphazard, slow response or none at all at a local council level. While it is true that local councils are not all influenced by the growth of these developments in the same way, a uniform guideline document or policy could assist towards more coherent responses. Provincial-level intervention can also include the development of a database with existing information on South Africa s gated communities (since many of the current types are a fairly recent phenomenon), examples of existing policies, problems and successes experienced by local councils, as well as lessons learnt. This could go along way to help those municipalities that are suddenly faced with the issue and have no experience in this regard, or do not have the capacity to commission in-depth research on the subject. Finally, a national policy that accords with the national planning policies could assist in establishing a common base for both provinces and local councils regarding this issue. It would also ensure that these types of developments occur in accordance with the main planning and urban development policies in the country, and that the vision and goals of these documents are therefore carried through at a local level when it comes to the development and establishment of different types of gated community in South Africa. A number of actions from different spheres of government are therefore called for. Although it is generally accepted that gated communities are a direct response to high crime rates, government should not adopt a laissez-faire attitude. While it is acknowledged that it is not an easy issue to deal with and often extremely contentious, intervention from at least local and often provincial government is required to give guidance. This is especially so in areas with a high demand for gated communities. Despite the limitations on what the state can do in terms of crime prevention, it still has a responsibility towards collective action where applicable. Many social problems and challenges facing local councils will not be solved on their own, and need stronger state intervention to guide future development towards the envisioned outcomes. This points towards two main areas of consideration when one considers the policy implications of gated communities and appropriate action. These two questions are: What are the main influences giving rise to the demand for these types of developments? CSIR Building and Construction Technology 75

83 What are the major impacts or implications of these types of developments likely to be in the future? In other words, decision-makers need to consider both the causes (drivers) of different types of gated community, as well as the consequences (outcome and impact) thereof in cities in the future. This would allow them to establish a balanced approach between immediate needs and long-term planning goals. 6.5 Directions for future study or research This study mainly focused on the current situation (location, extent and responses) regarding two types of gated communities in South Africa. It also highlighted a number of aspects likely to be influencing the demand for and thus distribution of these two types. The study did not, however, address the impact or implications of gated communities, which would constitute the second important driver for policy formulation. It is therefore imperative that future research include a focus on the impact and implications of different types of gated communities in the country as well. This will necessarily involve different methods to obtain the relevant information. CSIR Building and Construction Technology is currently conducting in-depth research of four case study areas in the cities of Johannesburg and Tshwane to determine, among other things, the future impact and implications of enclosed neighbourhoods and security estates on urban areas. While this may provide valuable information regarding the situation in Gauteng and allow for some generalisation, research has shown that these types of development can be very context-specific and differ substantially in different geographical areas. Local authorities may therefore have to consider this when dealing with these types, and may have to engage in their own research to establish a secure basis to support relevant decision-making based on current and up-to-date research. CSIR Building and Construction Technology 76

84 REFERENCES Lawrence, R. Comparing patterns of governance in Argentina, Brazil and South Africa, in Unisia Latin American Report 6(1), March Landman, K. (2000) An overview of enclosed neighbourhoods in South Africa. CSIR Publication: Pretoria. CSIR Building and Construction Technology 77

85 APPENDIX A: SURVEY QUESTIONNAIRE CSIR Building and Construction Technology 78

86 Section C: Luxury security villages or estates (please complete if you have any luxury security villages in your municipal area) Explanation: These large, luxury security villages include country and golf estates, as well as other mainly residential estates (smaller security villages are not included) 19. In which of the towns / metropolitan subcouncil or ward areas that fall under your municipality do luxury security villages occur? (Please name them) QUESTIONNAIRE ON GATED COMMUNITIES IN SOUTH AFRICA March 2002 No 0007 Name of municipality: Ekurhuleni Municipality Department: This questionnaire deals with physical areas that are fenced or walled off from their surroundings. Access to these areas are prohibited or controlled by means of gates or booms. Examples include enclosed neighbourhoods (through road closures) or security villages. Section A: General questions 1. What is the estimated population of your municipality? 20. In which types of urban areas do luxury security estates or villages, for example golf or country estates, mainly occur? (Please tick most appropriate one) Urban core (including CBD) Inner suburbs Peripheral suburbs/urban periphery Traditional townships 21. Approximately how many (estimated number) of these luxury security villages or estates are there in your municipal area? Additional information: Person who completed questionnaire: Contact phone number and/or address: Thank you very much for taking the time to complete this questionnaire. Please post it in the envelope provided or fax it to: (early response would be much appreciated and reduce the necessity for sending out reminders). For any queries, please contact: Karina Landman Tel: klandman@csir.co.za 2. Are there any existing neighbourhoods that have been closed off through road closures in your municipal area? Yes No 3. Are there any new security developments, such as secure office parks or luxury security villages (such as golf or country estates) in your municipal area? (Please tick appropriate ones) Security office parks Large residential security villages 4. Is there a tendency to secure existing townhouse complexes in your municipal area? (Please tick appropriate one) Yes No Don t know 5. Is there a tendency to secure high-rise apartment blocks in your municipal area? (Please tick appropriate one) Yes No Don t know

87 6. Has any form of enclosed area with controlled access been addressed as a major issue in your local urban planning policies? Yes No 7. Has any form of enclosed area with controlled access been addressed as a major issue in IDPs? Yes No 8. If yes, when was it included or when was the IDP prepared? Year Section B: Enclosed neighbourhoods (please complete if you have any form of enclosed neighbourhoods in your municipal area) Explanation: Enclosed neighbourhoods refer to the closure of existing neighbourhoods through booms and gates across roads. They are often referred to as road closures. 9. In which of the towns / metropolitan sub council or ward areas that fall under your municipality do enclosed neighbourhoods occur? (Please name them) 11. Do you have any policies in place on neighbourhood / road closures? (Please tick appropriate one) Yes No 12. If yes, when was it prepared? Year 13. If the answer to question 11 was no, are you in the process of developing or revising any policies on neighbourhood / road closures? Yes No 14. Have you received any applications for neighbourhood / road closures? (Please tick appropriate one) Yes No 15. If yes, how many applications have been received? 16. How many of these applications were approved? 17. How many road closures (physical structures across roads) in your municipal area? 10. In which types of urban areas do enclosed neighbourhoods mainly occur? (Please tick most appropriate one) Urban core (including CBD) Inner suburbs Peripheral suburbs/urban periphery Traditional townships 18. Are you aware of any illegal closures in your area? (Please tick appropriate one) Yes No

88 APPENDIX B: COVERING LETTER Date Attention: Municipality Corporate & Legal Department PO Box City/town Code CSIR Building and Construction Technology PO Box 395 Pretoria 0001 Tel: RE: QUESTIONNAIRE ON GATED COMMUNITIES IN SOUTH AFRICA Dear Sir: CSIR Building and Construction Technology has been conducting research on gated communities in South Africa for over two years. The aim of this ongoing research is to provide relevant and accurate information to assist with decision-making on gated communities in the country. At this stage we are particularly interested in the extent and location of these developments in the country. In order to determine this, I am sending the attached questionnaire to all the local municipalities in South Africa in the hope that we can establish more accurate data. Gated communities include security townhouse complexes, larger security estates, enclosed neighbourhoods, etc. This questionnaire deals with only two types of gated communities, namely: Enclosed neighbourhoods (road closures) These are existing neighbourhoods that have been fenced in through the closure of public roads. Large luxury security villages or estates Such as golfing or country estates and other large mainly residential security areas. (Smaller security townhouse complexes - with less than 50 units - are not included.) I will greatly appreciate it if the relevant person in your municipality can complete the attached short questionnaire. An executive summary of the report will be made available to all those municipalities that forward their completed questionnaires to the CSIR. Yours sincerely, Karina Landman Research Architect/Urban Designer CSIR Building and Construction Technology CSIR Building and Construction Technology 79

89 APPENDIX C: MAIN VARIABLES AND CODING Name Questionnaire number LA number LA name Estimated population Existing neighbourhoods Secure office parks Large residential security villages Closing of townhouse complexes Securing high-rise apartment blocks Urban policies Addressed in IDPs Added to IDPs year of IDP Types of urban areas in which enclosed neighbourhoods occur Policies on enclosed neighbourhoods Coding Source: Demarcation Board SA Real number 8 = not completed 9 = not applicable 1 = yes 2 = no 8 = not completed 9 = not applicable 1 = yes 2 = no 1 = yes 2 = no 8 = not completed 9 = not applicable 1 = yes 2 = no 3 = don t know 8 = not completed 9 = not applicable 1 = yes 2 = no 3 = don t know 8 = not completed 9 = not applicable 1 = yes 2 = no 8 = not completed 9 = not applicable 1 = yes 2 = no 8 = not completed 9 = not applicable Real number 8 = not completed 9 = not applicable 1 = urban core 2 = inner suburbs 3 = peripheral suburbs/ urban periphery 4 = traditional townships 5 = more than one 6 = rural 8 = not completed 9 = not applicable 1 = yes 2 = no 8 = not completed 9 = not applicable CSIR Building and Construction Technology 80

90 Name Year policy was prepared Developing policy Applications for neighbourhood enclosures How many applications How many applications approved Number of road Closures Illegal closures Types of urban areas in which large luxury security villages occur Number of large luxury security villages Local authorities per province % Urban population 1 = % 2 = 61 80% 3 = 41 60% 4 = 21 40% 5 = 0 20% Metro/ non-metro Coastal/ Inland Coding Real number 8 = not completed 9 = not applicable 1 = yes 2 = no 8 = not completed 9 = not applicable 1 = yes 2 = no 8 = not completed 9 = not applicable Real number 999 = not applicable 888 = not completed Real number 999 = not applicable 888 = not completed Real number 999 = not applicable 888 = not completed 1 = yes 2 = no 8 = not completed 9 = not applicable 1 = urban core 2 = inner suburbs 3 = peripheral suburbs/ urban periphery 4 = traditional townships 5 = more than one 8 = not completed 9 = not applicable Real number 999 = not applicable 888 = not completed 1 = Gauteng 2 = Limpopo Province 3 = Mpumalanga 4 = KwaZulu-Natal 5 = Western Cape 6 = Northwest Province 7 = Northern Cape 8 = Free State 9 = Eastern Cape 1 = metro 2 = non-metro (local municipality) 1 = coastal 2 = inland CSIR Building and Construction Technology 81

91 Name Popular tourist destinat. Characteristics and size of cities in municipal areas Coding 1 = popular tourist destinations (within municipal area) 2 = other (rest) 1 = metropolitan municipalities 2 = municipalities with large cities, including provincial capitals that are not part of metro municipalities 3 = municipalities with medium cities/larger towns 4 = municipalities with smaller towns and predominantly farmland (predominantly rural municipalities) CSIR Building and Construction Technology 82

92 APPENDIX D: CROSS-BOUNDARY LOCAL MUNICIPALITIES IN SOUTH AFRICA Table D1: Municipalities and CBLCs Provinces Number of municipalities Municipalities including CBLCs Eastern Cape Free State Gauteng GT/MP (CBLC) 1 GT/NW (CBLC) 1 KwaZulu-Natal Mpumalanga MP/LP (CBLC) 4 Northern Cape NC/NW 2 Limpopo North West Western Cape CSIR Building and Construction Technology 83

93 APPENDIX E: CROSS-TABULATION OF ENCLOSED NEIGHBOURHOODS WITH OTHER TYPES OF GATED COMMUNITIES IN SOUTH AFRICA Table E.1: Enclosed neighbourhoods and security office parks in all municipalities Existing closed neighbourhoods Yes No New security office park developments Not completed Not applicable Total Yes No Total Table E.2: Enclosed neighbourhoods and large security estates in all municipalities Existing closed neighbourhoods New large residential security estates Yes No Not completed Not applicable Total Yes 15 12,82 65,22 53, No Total CSIR Building and Construction Technology 84

94 Table E.3: Table of enclosed neighbourhoods and townhouse complexes in all municipalities Enclosed neighbourhoods Townhouse complexes Yes No Don't know Not completed Total Yes No Total Table E.4: Enclosed neighbourhoods and high-rise apartment blocks Enclosed neighbourhoods Yes No High Rise Apartments Don't know Not completed Not applicable Total Yes No Total CSIR Building and Construction Technology 85

95 APPENDIX F: CROSS-TABULATION OF ENCLOSED NEIGHBOURHOODS AND SECURITY ESTATES WITH PROVINCES IN SOUTH AFRICA Table F.1 Distribution of enclosed neighbourhoods in provinces Provinces Existing neighbourhoods that are closed off Yes No Total Eastern Cape Free State Gauteng KwaZulu-Natal Mpumalanga MP/NP (CBLC) Northern Cape NC/NW (CBLC) Northern Province (Limpopo) CSIR Building and Construction Technology 86

96 Table F.1 Distribution of enclosed neighbourhoods in provinces Provinces Existing neighbourhoods that are closed off Yes No Total North West Western Cape Total Table F.2: Number of municipalities with large security estates per province Large security estates Provinces Yes No Not Completed Not Applicable Total Eastern Cape Free State Gauteng KwaZulu-Natal Mpumalanga CSIR Building and Construction Technology 87

97 Table F.2: Number of municipalities with large security estates per province Large security estates Provinces Yes No Not Completed Not Applicable Total MP/NP (CBLC) Northern Cape NC/NW (CBLC) Northern Province North West Western Cape Total CSIR Building and Construction Technology 88

98 APPENDIX G: LOCATION OF SECURITY ESTATES IN THE CITY OF CAPE TOWN Prepared by the City of Cape Town to accompany the questionnaire information July 2002 CSIR Building and Construction Technology 89

99 CSIR Building and Construction Technology 90