The suffocating cost of transport in the Gauteng City-Region

Introduction

Where one lives (and in turn where one works or does shopping) fundamentally shapes how much one spends on transport. In post-apartheid South Africa, the household cost of transport has long been recognised as a critical policy issue. High transport costs are a key spatial justice concern in that they directly influence access to urban opportunity and amenity, and in turn socio-economic mobility.

Given the country’s high dependence on public transport among lower-income households, the South African government has an existing policy target, articulated in the 1996 White Paper on National Transport Policy, to ensure that household expenditure on public transport is limited to no more than 10% of the total household income (Department of Transport, 1996).

Of course, some have argued that the target seems overly ambitious given that globally transport expenditure sits between 10-15% (Knipe and Krygsman, 2024). Others have pointed out that the exact grounds for the 10% target are not clear (Behrens and Venter, 2005). Nevertheless it remains an important benchmark, and whether or not this target is being realised has been analysed in various processes and through various datasets over the years.

For example, Kerr (2021) notes that in 1995 only 11% of households spent more than 10% of their income on transport. By 2015, this had increased to 46% (Kerr, 2021). In a 2020 investigation the Competition Commission (2020:1) highlighted that “South Africans spend a significantly high proportion of disposable income on public transport (over 20%) against a benchmark of 10% for developing countries. Over 73% of rural workers spend more than 20% of their monthly household income per capita on public transport, while in urban areas the percentage is 60.1% and in metros 54.7%.” And according to a recent report by Statistics South Africa (StatsSA, 2024), total transport expenditure now represents 15.3% of household income nationally. The report spells out that while private car users spend the most in absolute terms, public transport users, and particularly minibus taxi users, spend a greater share of their income on travel. Car users tend to belong to wealthier households, resulting in a smaller proportional expenditure despite higher nominal spending. Given continued fuel price inflation, weak growth in wages, and the continued absence of subsidies for minibus taxis, the proportion is almost certainly likely to rise still higher in the future.

While these analyses have put a spotlight on the challenge in overall terms, the spatial dimensions of expenditure on transport have rarely been visualised. The map presented here provides a spatialised analysis of the share of per capita household income being spent on transport expenditure. It reveals a deeply inequitable geography of transport expenditure in the Gauteng City-Region, with residents in many of Gauteng’s townships and on its urban periphery facing financially and economically suffocating transport costs, often exceeding 50% of household income per capita.

Understanding the data

This Map of the Month uses data from the GCRO’s Quality of Life survey 7 (2023/24) (QoL 7) (see methods note below for more detail on the QoL 7 questions used). It presents calculations, averaged per ward, of the percentage of total monthly household income per capita that respondents are spending on transport.

South Africa’s national policy target of no more than 10% spend on transport is based on monthly household expenditure and income. However, the Quality of Life Survey does not collect information on the transport activity of each and every household member, so we cannot derive an aggregate household-level transport spend. An appropriate proxy is therefore to divide the individual respondent’s stated expenditure on transport by the total household income shared between all household members. We do this by dividing total monthly household income by the number of household members to get an estimate of per capita income. Per capita household income offers a more nuanced view of expenditure capacity, capturing how much income is available for the needs of each person in the household – food, clothing, healthcare, education, service costs, debt repayment, savings, and so on – that must be met each month.

To illustrate, consider two households of Lisa and Thabo:

• Lisa’s household earns the same total income as Thabo’s household. Both use minibus taxis to get to work.
• Lisa lives with five other people (a household of six) while Thabo lives with two (a household of three).
• Thabo therefore has roughly twice the per capita income capacity of Lisa. Even if both spend similar amounts on minibus taxi fares, Lisa’s transport expenditure represents a far higher proportion of the income available in her household.

In this Map of the Month we first calculate the monthly household income per capita of each QoL survey respondent; we then work out the share of this income per capita going to transport expenditure; and finally we tally the resulting percentages and average at the ward level to get a spatial perspective.

This is shown in Map 1, which gives the average percentage per ward of households’ per capita monthly household income that individual respondents incur on transport. Subsequent Maps 2 and 3 nuance the picture, distinguishing between those respondents whose most frequent trip is to work and those whose most commonly made trip is to shop.

Spatial differentiation in the share of household income per capita used for transport expenditure

Map 1: Percentage of household income per capita used for total transport expenditure across Gauteng.

In overall terms, GCRO’s QoL 7 (2023/24) data indicates that Gauteng residents spend on average some 29% of their household income per capita on transport, regardless of where they are travelling to, or how. However, as illustrated in the map, the percentage varies considerably between parts of the city-region. It is clear that in centrally located wards – which tend to have higher household incomes and that are close to places where employment, education or retail opportunities concentrate – people spend on average within 10-20% of their per capita household income on transport. However, transport costs rise sharply with distance from economic centres. Those living in township areas or on the outskirts of the city region, especially where commuting is reliant on minibus taxis, often face a much higher proportional burden. The challenge is particularly severe in the south of Johannesburg, in townships like Sebokeng, Vosloorus and Mamelodi, and the far northern and western edge of the city-region.

Where the cost of transport exceeds 40% or even 50% of per capita household income – five times the national policy target of 10% – it goes without saying that there will be far less money available for other monthly household essentials, not to mention savings and investments that are core to wealth building and economic prosperity.

Map 2: Percentage of household income per capita used for transport expenditure for work across Gauteng.

Map 3: Percentage of household income per capita used for transport expenditure for shopping across Gauteng.

When expenditure is further disaggregated into work and shopping trips, the spatial trend remains largely the same, although there are some differences. All three maps indicate clear hotspot areas like Shoshanguve (A) in the north, Sebokeng (N) in the south and Springs/KwaThema (L) in the east. When the data for those whose most frequent trip is to work is viewed separately, peripheral wards in the far north, far south and east of the city-region stand out, highlighting the high cost of work trips in particular for households on the outskirts commuting to economic centres.

Disaggregation by transport mode, race and sex

Figures 1 and 2 below should be read together, with both illustrating that the largest share (38%) of people travelling in Gauteng on a daily basis rely on minibus taxis for the longest part of their trip. The next largest proportion of commuters (27%) travel by private car as a driver. Those who mainly travel by minibus taxi spend 33,2% of their per capita monthly household income on transport. Notwithstanding the high cost of fuel, and the expense of maintaining a private vehicle, the proportion of monthly household income going to transport is lower for those who drive, at 32,8%, than it is for those who take taxis as their main mode. The clear and alarming conclusion is that minibus taxi use, the dominant public transport mode among Quality of Life survey respondents in Gauteng, is driving up the proportion of household income that must be devoted to transport expenditure.

Figure 1. Diagrammatic representation of the proportion of commuters using each transport mode for the longest part of their trip (the larger the block, the bigger the share of trips made using that mode).

Figure 2. Average transport expenditure as a percentage of monthly household income per capita, by transport mode for the longest part of the trip, race and sex.

Figure 2 further shows that African respondents spend on average 29,5% of their monthly household income per capita on transport, compared to just 23,2% for white respondents. Women spend 29,8% compared to 28,3% for men.

These patterns of transport affordability need to be understood as the structural outcome of South Africa’s spatial form. The persistent spatial mismatch between where low-income households live, and where jobs, schools, retail centres and other urban amenities are located, continues to drive up transport costs (Turok, 2012).

The spatial inequality inherited from apartheid, along with decades of post-apartheid housing delivery, has reinforced peripheral settlement patterns with large-scale RDP developments often situated on cheap land far from economic centres and public transport nodes. With limited coordination between land use planning and transport investment, our cities have failed to promote densification along existing corridors (SACN, 2016). The result is that poorer households are spatially taxed, paying both financially and in time for the distance to opportunity (Todes and Turok, 2017). Unless these structural linkages are addressed, transport interventions alone are likely to have limited impact.

Conclusion

Three key factors shape the patterns revealed by this map: household income, household size, and transport expenditure. Together, they expose the unequal geography of mobility in Gauteng, where households in townships and peripheral areas endure the heaviest transport burdens.

Unlike food costs, which remain relatively equivalent regardless of location in an urban area, transport costs scale with distance. For those living far from economic centres, each additional kilometre travelled represents a deepening financial constraint.

Given the post-apartheid imperative of providing more equitable and affordable access, the map highlights the need for the government to seriously consider a revised subsidy policy that reduces the cost burden for minibus taxi users, who account for the highest volume of public transport users and who are facing the highest relative cost burden. However, transport interventions alone are not sufficient. Without spatially targeted interventions, peripheral commuters will remain financially and economically suffocated by the cost of movement in the GCR. More needs to be done in terms of co-ordinated land-use and transport planning to shift the prevailing unequal spatial patterns in the GCR.

Note on data and methods

This analysis was based on data from the GCRO’s Quality of Life Survey 7 (2023/24) dataset, specifically questions 5.10 (“Approximately how much do you personally spend in total every month on transport?”) and 15.3 (“What is the total amount of money brought into the household per month by all household members?). Since question 15.3 was structured using income ranges, the midpoint value for each category was calculated to infer the total monthly household income. To normalise the household income by the number of people in the household, question 14.5 (“How many people, including you, live in this household?”) was incorporated in the analysis. Total household income was divided by the number of household members to ascertain the household income per capita. To derive the percentage of household income per capita going to transport for the relevant respondent, the responses to question 5.10 were divided by the values from question 15.3 (per capita) and then multiplied by 100. The result was a percentage of household income per capita expenditure spent on transport at the individual respondent level. Percentages were then tallied and averaged at the ward level.

For distinguishing transport expenditure percentages for work and shopping, question 5.3 (“Think about the trip that you make most often, from this dwelling. What is the purpose of this trip?”) was used to filter the respondent transport expenditure calculations (“1” for work and “4” for shopping) prior to averaging at the ward level. Specifically for the work and shopping transport expenditure maps, wards with less than 3 respondents were greyed out as it was deemed that there were not enough respondents to infer a ward level trend.

Finally for analysis per transport mode, racial population group and gender, questions 5.7 (“Last time you made this trip, what mode of transport did you use to cover the longest distance?”), A1 (“To which population group does the respondent belong?”) and A2 (“What is the sex of the respondent?”) were used to intersect the percentage transport expenditure calculations per respondent and averaged per question.

Throughout the analysis unweighted QoL data was used, as weighting significantly reduces the number of responses counted in the dataset in smaller peripheral wards.

References

Behrens, R. and Venter, C. (2005). Is the 10% policy benchmark appropriate? Proceedings of the 24th Southern African Transport Conference. Pretoria.

Competition Commission. (2020). Public Passenger Transport Market Inquiry: Provisional Findings and Recommendations Report. Pretoria: Competition Commission of South Africa.

Department of Transport. (1996). White Paper on National Transport Policy. Pretoria: Government of the Republic of South Africa.

Knipe, A. and Krygsman, S. (2024). The opportunity cost of household transport expenditure in South Africa. Journal of Transport and Supply Chain Management, 18. AOSIS.

SACN (South African Cities Network). (2016). State of Cities Report 2016: Cities as Effective Drivers of Local and National Development. Johannesburg: SACN.

Statistics South Africa (Stats SA). (2024). Income and Expenditure Survey 2022/2023: Statistical Release P0100. Pretoria: Statistics South Africa. Available at: https://www.statssa.gov.za/?p=18891 (Accessed: 28 October 2025).

Statistics South Africa (Stats SA). (2022). National Household Travel Survey 2021: Statistical Release P0320. Pretoria: Statistics South Africa. Available at: https://www.statssa.gov.za/?p=15341 (Accessed: 28 October 2025).

Sutherland, A., Kerr, A. (2021). The Evolution of Household Transport Spending in Post-Apartheid South Africa: 1995-2015. Cape Town: Southern Africa Labour and Development Research Unit, University of Cape Town. (SALDRU Working Paper Number 287).

Todes, A. and Turok, I (2017). Spatial inequalities and policies in South Africa: Place-based or people-centred? Progress in Planning, 118, 1–31.

Turok, I. (2012). Urbanisation and Development in South Africa: Economic Imperatives, Spatial Distortions and Strategic Responses. London: Human Settlements Group, IIED.

Cartography/mapping: Dr Laven Naidoo and Christian Hamann

Inputs, edits, and comments: Graeme Götz, Christian Hamann and Dr Samkelisiwe Khanyile

Suggested citation: Bickford, G., and Naidoo, L. (2025). The suffocating cost of transport in the Gauteng City-Region. GCRO Map of the Month, October 2025. Gauteng City-Region Observatory, Johannesburg. DOI: https://doi.org/10.36634/AWYS6444.