Mapping perceptions of (un)safety across the Gauteng City-Region

Introduction

South Africans are acutely aware of their safety and security in a way that is uncommon among residents of other stable, middle-income democracies. An enduring trend of the post-apartheid period has been incredibly high levels of violent crime. Over the past decade, the country has witnessed a steady rise in murders, peaking at 45 deaths per 100 000 people in 2023 – representing a 53% escalation rate from its lowest recorded level in 2012 (Bruce, 2023). Alongside this is the “pandemic” of gender-based violence (GBV) and femicide (Ramaphosa, 2024). The rate at which women are killed by intimate partners in South Africa is five times higher than the global average (Govender, 2023). An astonishing 36% of South African women report having experienced physical and/or sexual violence in their lifetime (Human Science Research Council, 2024). GBV is rooted in patriarchal gender norms, which both reinforce women’s subordinate status in society and pit men against one another in a hierarchy of masculinity, often manifesting in violence experienced at the hands of other men. Although South African men are the main perpetrators of GBV, it is important to acknowledge that men and boys are also victims of sexual violence (Human Sciences Research Council, 2024), and are more likely to experience physical violence in their lifetime than women (Brankovic, 2019).

Gauteng has a reputation for high levels of violent crime and interpersonal violence. Mkhize and Booysen (2024) use GCRO’s Quality of Life (QoL) Survey 7 (2023/24) data to demonstrate that a third of all respondents experienced some form of violence in the year before they were interviewed. However, people’s actual experience of crime, and their perceptions of personal safety and security, are not equivalent across all parts of the province or across all groups.

Literature links the occurrence of violence and crime in urban environments with various socio-economic factors, such as concentrated poverty, high unemployment, a lack of social mobility, distrust in authorities, and social and political disempowerment, particularly among young men (Salahub, Gottsbacher and de Boer, 2018). Similarly, research shows that in urban settings people’s perception of their exposure to the risk of violence varies greatly by economic status. Perceptions of higher vulnerability to crime are associated with various disadvantages such as poverty and income inequality (Kujala et al., 2019). So poor households are more likely to live in disadvantaged neighbourhoods that they perceive to be unsafe (Alloush and Bloem, 2022). Race is also a factor in perceptions of safety. Although crime affects all races in South Africa, a persistent myth suggests that white South Africans are disproportionately targets of violent crimes committed by black individuals (Kynoch, 2014). However, research indicates that black South Africans are more likely to fall victim to crime as they are unable to afford the security measures that many white South Africans can access, largely due to the privileges stemming from historical injustices and inequalities (Silber and Geffen, 2009; Mulamba, 2024).

Drawing on the GCRO’s QoL Survey 7 (2023/24), this Map of the Month shows clear spatial disparities in the perceived lack of safety – (un)safety – across wards in Gauteng. Our further analysis below digs deeper to explore the relationship between (un)safety and various factors: sex, race, age, income and dwelling type.

Perceptions of (un)safety across the GCR

The GCRO’s QoL 7 survey (2023/24) asked a range of questions on experiences of crime and violence and perceptions of safety. For this Map of the Month, an (un)safety index was developed for each respondent derived from responses to three questions: “How safe do you feel walking in the area where you live during the day”; “How safe do you feel walking in the area where you live after dark?”; and “How safe do you feel at home?”. For each of these questions, respondents were asked to rank how safe they felt on a likert scale from 1 to 5, with 1 being very safe and 5 being very unsafe. A total safety score for each respondent was calculated by adding together the numerical value associated with each response category, and dividing that total by the number of questions (3) to get the average safety score – an (un)safety index – per respondent. For our spatial analysis we calculated the average safety score per ward by adding up all respondent’s (un)safety index scores and dividing that by the total number of respondents per ward.

We acknowledge that deriving a general safety score in this way presents some limitations, the first being that responses for each safety question are equally weighted when determining the (un)safety index, even though the importance that each respondent accords to each of the safety questions may differ depending on context. For instance, a woman could feel very safe walking in her neighbourhood in the day, but feel very unsafe in her home, and may therefore weigh this latter variable as much more important. Further, a ward level score may obscure significant variance within wards.

Figure 1 illustrates the spatial variability of perceived (un)safety across Gauteng wards. The wards with the highest scores are those that are perceived to be most unsafe by their residents. The resultant map uses a graduated scale from white to dark purple, corresponding to a safety score from 1 (wards perceived as very safe) to 5 (wards perceived as very unsafe). Importantly, no wards recorded a score of very safe, and therefore there are no white wards reflected on the map. Most of the wards in the province fall in the ranges of ‘safe’ or ‘neither safe nor unsafe’. The northern suburbs of the City of Johannesburg, as well as the southern suburbs of the City of Tshwane, mostly fall into these categories. The map marks with a green border the five wards perceived by their residents to be the safest in the province: Mooikloof Ridge, Die Wilgers, Midvaal, Merafong City and Hillshaven. Notably, the two safest wards, both with safety scores of 1.7, Mooikloof Ridge and Die Wilgers, are gated communities in the City of Tshwane.

Figure 1: Perceptions of (un)safety across wards in Gauteng mapped using the average ward safety score ranging from very safe to very unsafe. The wards highlighted in green are the five wards perceived to be most safe. Labels F-G correspond to the top five safest wards. The wards highlighted in yellow are those perceived to be most unsafe. Labels A-E correspond to the top five most unsafe wards. Source: GCRO QoL 7 Survey (2023/24)

Perceptions of unsafety tend to be higher in areas marked by deprivation, poverty, family disruption and residential instability (Meth, 2017). The map reveals that areas perceived to be unsafe in Gauteng tend to be informal settlements either surrounding large townships or adjacent to key economic hubs and developments, or informal settlements and/or townships in rural areas on the periphery of the province. Further, small pockets of unsafety are also evident within inner-city areas. This spatial pattern reflects the distribution of multidimensional poverty in the province, which is concentrated in areas of limited economic activity situated on the outskirts of the province, as well as in clusters in the heart of metropolitan areas (Mushongera et al., 2015).

A line of wards on the rural periphery in Merafong City, and on the edge of the City of Tshwane, have notably lower levels of perceived safety. Similarly, there are pockets of unsafety in the urban core and in areas south of Johannesburg, particularly those surrounding Soweto, as well as in parts of the City of Ekurhuleni. There are 127 wards with an index score above 3 (very unsafe). The map marks with a yellow border the five wards with the highest scores, and which are therefore the most unsafe: Holomisa Section, Kokosi Extension 5, Leratong Village, Westonaria and Claremont. Notably, the two wards perceived to be least safe, Holomisa Section in Rand West and Kokosi Extension 5 in Merafong City, are both characterised by elements of informality.

It is worth noting that some of the wards that are perceived to be safe border other wards perceived to be unsafe or very unsafe. Studies such as Thornton et al. (2023) show that the real and perceived threat of crime (particularly burglary) in high-income areas results in residents turning to physical barriers – higher walls, home security systems and private armed security – to ward off the risk of crime and create a heightened sense of safety. This, in turn, embeds inequality in space, as residents of lower-income neighbourhoods are deterred, and often physically barred, from safer communities (World Bank, 2011). This can exacerbate already heightened experiences of insecurity within adjacent poorer neighbourhoods (Koonings and Kruijt, 2007).

Perceptions of safety across Gauteng municipalities

Figure 2 shows the mean safety scores across municipalities in Gauteng and the average scores of each ward in each municipality corresponding to groupings of 'safe', 'neither safe nor unsafe', and 'unsafe'. Merafong City reports the highest mean (un)safety score at 3, although the average (un)safety index of its wards reflect significant variability (1.8 - 3.8), indicating a combination of both very safe and very unsafe areas within the municipality. A similar pattern is observed in Rand West, which has a mean (un)safety score of 2.9. The City of Johannesburg, Emfuleni, Mogale City and the City of Ekurhuleni all have a mean (un)safety score of 2.8, with some very unsafe areas.

On the lower end of the spectrum, the Midvaal and Lesedi municipalities have the lowest mean (un)safety score at 2.4, with the index ranging from 1.8 - 3. The highest ward-level average scores in both remain modest when compared to those for other municipalities, suggesting that the two municipalities are generally perceived as safer by their residents.

The City of Tshwane, with a mean (un)safety score of 2.6, occupies a middle position, but stands out for the wide range in its ward-level average scores. Its highest (un)safety score is 3.8, while its lowest score is 1.7. This reflects the contrasting perceptions of safety in different areas within the metro, with some wards seen as considerably safer than others.

Figure 2: (Un)safety index of respondents and average safety score across Gauteng municipalities. (circle: (un)safety index corresponding to safe; square: (un)safety index corresponding to neither safe nor unsafe; and triangle: (un)safety index corresponding to unsafe. Black dot: mean safety score per municipality). Source: GCRO QoL 7 Survey (2023/24).

Perceptions of (un)safety across demographic and socio-economic factors

The (un)safety index per respondent was further investigated against socio-economic and demographic factors, specifically gender, race, age and income (Figure 3). Overall, more than half (51%) of the respondents in Gauteng fall in the band of index scores considered neither safe nor unsafe, while 22% feel safe and 27% feel unsafe. This trend is consistent across genders, with 51% of both male and female respondents expressing that they feel neither safe nor unsafe. However, only 18% of women feel safe, compared to 26% of men. Conversely, 31% of women feel unsafe, compared to 23% of men.

A similar distribution in perceptions of safety is evident across age groups. Overall, around a fifth of respondents in each age group report feeling safe, while about a quarter in each group report feeling unsafe. Notably, middle-aged respondents (age groups of 35-44, 45-54 and 55-64) exhibit the highest proportion of individuals who feel unsafe, with 28% in each of these groups reporting a sense of unsafety.

When analysed by racial group, the data show that white respondents report the highest sense of safety, with 29% feeling safe and only 19% feeling unsafe. In contrast, 20% of black African respondents feel safe, while 29% feel unsafe. Similarly, 21% of Indian/Asian respondents report feeling safe, whereas 27% feel unsafe. Notably, the coloured group has the highest proportion of respondents who feel unsafe, with only 19% reporting feeling safe and 33% reporting feeling unsafe.

Consistent with the existing literature (Meth, 2017), our results show that residents residing in informal housing report higher levels of insecurity compared to those in formal housing. Among respondents residing in formal dwellings, 26% reported feeling unsafe, while 23% feel safe. In contrast, 35% of those living in informal dwellings expressed feeling unsafe, with only 14% feeling safe. Regardless of housing type, 51% of respondents in both formal and informal dwellings indicated that they felt neither safe nor unsafe (Figure 3).

Figure 3: Perceptions of (un)safety by demographic and socio-economic factors. Source: GCRO QoL 7 Survey (2023/24).

Figures 3 and 4 show an increase in perceived safety as monthly income increases. Among respondents in the lowest income bracket (R1 - R800 per month), only 16% report feeling safe, while 39% feel unsafe. A modest but consistent rise in perceived safety is observed across income groups. Notably, there is a six-percentage-point increase in safety perceptions between those earning R3 201 - R12 800 per month (18% feel safe) and those in the R12 801 - R25 600 bracket (24% feel safe). This trend continues at higher income levels, with each successive bracket showing a comparable increase in the proportion of individuals who feel safe. A significant shift occurs at the R25 601 - R51 200 income range, where the proportion of respondents feeling unsafe drops sharply to 21%, compared to approximately a third in the lower three brackets. The highest level of perceived safety is recorded among respondents earning R102 401 - R204 800 per month, where nearly half (49%) report feeling safe, and only 9% feel unsafe. However, in the highest income groups perceptions of safety do decline and feelings of insecurity are elevated.

Further analysis, shown in Figure 4, reveals a strong positive correlation between income and perceived safety exists in Gauteng, as indicated by the linear regression equation (y = 2.3132x + 9.718) and correlation coefficient (R2 = 0.82). The analysis shows that as income increases, the perception of safety improves. In contrast, there is a negative correlation between income and perceptions of unsafety among respondents, as indicated by the linear regression equation (y = -1.8495x + 39.534) and the correlation coefficient (R2 = 0.79). The relationship reiterates that there are heightened perceptions of unsafety in respondents who earn lower incomes (Figure 4).

Figure 4: Correlation between income and perceived (un)safety in Gauteng. Source: GCRO QoL 7 Survey (2023/24).

Conclusion

This Map of the Month reveals clear spatial disparities in perceived (un)safety across Gauteng. There is a strong correlation between higher income and perceived safety, and hence the spatial distribution of safety perceptions aligns with broader socio-economic patterns. Affluent, suburban and gated communities report higher levels of safety, while informal settlements, low-income townships, pockets within inner-cities and areas on the rural periphery experience a heightened sense of insecurity. White residents, who are typically more affluent, report the highest sense of safety, while coloured residents feel the most unsafe. Gender differences are also pronounced, as women consistently feel more unsafe than men. The analysis thus confirms the link established in the literature between wealth and safety, and underscores the intersectionality of income, gender, race and perceived security.

References

Alloush, M. and Bloem, J.R. (2022) ‘Neighborhood violence, poverty, and psychological well-being’, Journal of Development Economics, 154, p. 102756.

Brankovic, J. (2019) What Drives Violence in South Africa. Research Brief. Centre for the Study of Violence and Reconciliation. Available at: https://csvr.org.za/pdf/What-Drives-Violence-in-South-Africa.pdf (Accessed: 24 February 2025).

Bruce, D. (2023) Analysis of murder in SA’s deadliest provinces calls for policing rethink, ISS Africa. Available at: https://issafrica.org/about-us/press-releases/analysis-of-murder-in-sas-deadliest-provinces-calls-for-policing-rethink (Accessed: 24 February 2025).

Galtung, J. (1969) ‘Violence, Peace, and Peace Research’, Journal of Peace Research, 6(1), pp. 167–191.

Govender, I. (2023) ‘Gender-based violence – An increasing epidemic in South Africa’, South African Family Practice, 65(1). Available at: https://doi.org/10.4102/safp.v65i1.5729.

Human Sciences Research Council (2024) The First South African National Gender-Based Violence Study: A Baseline Survey on Victimisation and Perpetration. Research Report. Cape Town: Human Sciences Research Council. Available at: https://hsrc.ac.za/wp-content/uploads/2024/11/Executive-Summary-Final-16-November-2024-1.pdf.

Koonings, K. and Kruijt, D. (eds) (2007) Fractured cities: Social exclusion, urban violence and contested spaces in Latin America. Zed Books. Available at: https://doi.org/10.5040/9781350220225.

Kujala, P., Kallio, J. and Niemelä, M. (2019) ‘Income Inequality, Poverty, and Fear of Crime in Europe’, Cross-Cultural Research, 53(2), pp. 163–185. Available at: https://doi.org/10.1177/1069397118799048.

Meth, P. (2017) ‘Informal housing, gender, crime and violence: The role of design in urban South Africa’, British Journal of Criminology, 57(2), pp. 402–421.

Mkhize, S.P. and Booysen, F. (2024) Experiences of violence, crime and safety: Findings from the GCRO’s Quality of Life survey 7 (2023/24). Data Brief. Gauteng City-Region Observatory. Available at: https://doi.org/10.36634/XRBF4904.

Mulamba, K.C. (2024) ‘Modelling the relationship between living in informal dwellings and fear of crime in Gauteng Province, South Africa’, Journal of Urban Affairs, pp. 1–20. Available at: https://doi.org/10.1080/07352166.2024.2388048.

Mushongera, D., Ngwenya, P. and Zikhali, P. (2015) A multidimensional poverty index for Gauteng (GMPI). Gauteng City-Region Observatory. Available at: https://doi.org/10.36634/LBXG2515.

Nixon, R. (2011) Slow Violence and the Environmentalism of the Poor. Harvard University Press.

Ramaphosa, C. (2024) President Cyril Ramaphosa: National Council on Gender-Based Violence and Femicide Bill and the National Prosecuting Authority Amendment Bill | South African Government. Available at: https://www.gov.za/news/speeches/president-cyril-ramaphosa-national-council-gender-based-violence-and-femicide-bill (Accessed: 2 February 2025).

Salahub, J.E., de Boer, J. and Gottsbacher, M. (2018) ‘Global South theories of urban violence, poverty and inequalities’, in J.E. Salahub, J. de Boer, and M. Gottsbacher (eds) Social Theories of Urban Violence in the Global South: Towards Safe and Inclusive Cities. Abingdon, Oxon: Routledge, pp. 1–12.

Silber, G. and Geffen, N. (2009) ‘Race, class and violent crime in South Africa: Dispelling the “Huntley thesis”’, South African Crime Quarterly, (30), pp. 35–43. Available at: https://doi.org/10.17159/2413-3108/2009/i30a897.

Thornton, A. et al. (2023) ‘Crime, Income, and Inequality: Nonlinearities under Extreme Inequality in South Africa’, Economic Development and Cultural Change, 72(1), pp. 207–240. Available at: https://doi.org/10.1086/719646.

World Bank (2011) Violence in the City: Understanding and Supporting Community Responses to Urban Violence. Research Report 63888. Washington, DC: The World Bank. Available at: https://documents1.worldbank.org/curated/en/524341468331181450/pdf/ 638880WP0Viole00BOX361532B00public0.pdf#page=32.19 (Accessed: 24 February 2025).

Edits and inputs: Graeme Götz, Christian Hamann, Samkelisiwe Khanyile, Laven Naidoo

Suggested citation: Singh, S. & Arnold, S. (2025). Mapping perceptions of (un)safety across the Gauteng City-Region. GCRO Map of the Month, Gauteng City-Region Observatory, February 2025.