Africa's two-tailed economy. South Africa's share of people living above US$15 a day is higher than in all but five of the 49 African countries we cover — and more than one in four South Africans live below US$3.65. Population 63.1 million (Stats SA, mid-2025); GDP US$427.1B in 2025, growing 1.1% in real terms, with inflation at 3.2%.
South Africa in 2025 is a US$427.1 billion economy with 63.1 million people (Stats SA mid-year estimate), growing at 1.1% in real terms. GDP per capita is US$6,580 in current US dollars and US$15,871 on a PPP basis. The urbanisation share is 63.7%, up from 63.2% in 2015.
South Africa has 8.9 million households that are middle class or above in 2025, 47.8% of all households. The functional urban areas of Johannesburg, Cape Town, Tshwane and Durban hold 3.6 million of them — 40.9% of the national total — and Johannesburg's alone holds 1.8 million, about one in five. Gqeberha is the fifth metro market, at 161k. Johannesburg's functional urban area spans 14 built-up clusters, which is why its 11.5 million people exceed any single municipality's population. Other cities in our data include Polokwane and Bloemfontein.
| Functional urban area | Population 2025 | Middle-class households 2025 | Density |
|---|---|---|---|
| Johannesburg | 11.5M | 1.8M | 55.1% |
| Cape Town | 5.0M | 762k | 53.3% |
| Tshwane | 3.7M | 541k | 51.2% |
| Durban | 3.3M | 510k | 53.4% |
| Gqeberha | 1.1M | 161k | 50.2% |
Middle-class households are households in Pan Africa Data's middle-class-or-above classes (above US$5.50 per person per day, 2017 PPP). Density is the share of a city's households in those classes. Figures rounded. Five selected cities of the 17 urban areas we cover in South Africa. Functional urban areas are built-up settlement clusters defined by the UN DEGURBA method, not municipal boundaries, so they do not match metro municipality populations. Pan Africa Data applies South Africa's national Gini coefficient within each urban area, so these figures size each urban area's middle class; they do not measure inequality between or within cities.
In December 2025 Statistics South Africa published new poverty lines and headcounts from its 2022/23 Income and Expenditure Survey. Depending on which of its three lines is used, 17.6%, 37.9% or 66.7% of South Africans were poor in 2023. Pan Africa Data's marginalised class — below US$3.65 a day on the 2017 PPP basis — held 38.4% of South Africans in 2025.
The closeness of 38.4% to 37.9% is a coincidence of two different lines, not agreement. Stats SA's lower-bound line of R1,300 a month is worth about US$4.81 a day in 2017 PPP terms, well above our US$3.65 threshold. At the same US$3.65 line, Stats SA's 2023 figures imply roughly 26% and the World Bank reports 30.4%, against Pan Africa Data's 35.8% for that year. For South Africa our marginalised share therefore runs about 5 to 10 percentage points high. We disclose this on our validation page, and a calibration refinement in the next model version is expected to close most of it. For South African policy questions, use Stats SA's lines. For comparing South Africa with Kenya, Egypt or Nigeria, use a common international line, and read our South African figure as an upper estimate.
Pan Africa Data publishes population shares across five fixed-boundary income classes on the 2017 PPP basis: marginalised (below $3.65/day), low ($3.65–5.50), middle ($5.50–15), upper-middle ($15–46.52), and high (above $46.52). The table shows South Africa's national distribution for 2015, 2020, 2024 and 2025. Annual history from 2000 and forecasts to 2035 are in the full dataset.
| Year | Marginalised | Low | Middle | Upper-middle | High | Total pop |
|---|---|---|---|---|---|---|
| 2015 | 42.3% | 14.7% | 28.3% | 12.7% | 2.1% | 56.7M |
| 2020 | 38.4% | 14.6% | 29.8% | 14.8% | 2.4% | 60.6M |
| 2024 | 37.2% | 14.6% | 30.3% | 15.4% | 2.5% | 64.0M |
| 2025 | 38.4% | 13.8% | 28.9% | 16.2% | 2.7% | 64.9M |
Each class is rounded independently, so the columns do not always sum exactly to the combined figures quoted in the text. The Total pop column is Pan Africa Data's population series, which follows World Bank and United Nations estimates so that figures stay comparable across all 54 countries we cover. Statistics South Africa's mid-2025 estimate, quoted in the summary above, is 63.1 million, 2.9% lower. The shares do not depend on which population total is used.
Gini coefficient in 2025: 62.1. The same national figure is applied to every South African city and locality in Pan Africa Data.
Every number on this page has been pulled from Pan Africa Data's platform and reconciled against source. Historical population, GDP and inflation come from World Bank World Development Indicators; 2025 estimates draw on the IMF World Economic Outlook (April 2026) and the World Bank Macro Poverty Outlook (April 2026). Urban population share and city envelopes come from UN World Urbanization Prospects 2025. Functional-city populations follow the UN WUP 2025 DEGURBA cluster methodology.
Pan Africa Data's income distribution model is calibrated against the World Bank Poverty and Inequality Platform at each country's income-group poverty threshold. Independent validation of the calibration against six African national statistics offices is published at panafricadata.com/validation. Full model methodology is at panafricadata.com/methodology. Model version referenced throughout this factbook is PAD-v1.2-2026.
Poverty lines and headcounts: Statistics South Africa, Poverty Trends in South Africa: An examination of absolute poverty between 2006 and 2023 (11 December 2025), based on the Income and Expenditure Survey 2022/23. Population: Statistics South Africa mid-year population estimates, 2025. Pan Africa Data's own population series (64.9 million in 2025) follows World Bank and United Nations estimates; the gap with Stats SA's 63.1 million is under 3%. Cross-country comparisons use Pan Africa Data's 2025 figures for the 49 African countries with income-class data.
This page shows 5 of the 17 urban areas and none of the 84 localities we cover in South Africa, one combined income class, 2025 only. The full dataset has every city and locality across all five income classes, annually from 2000 with forecasts to 2035, in PPP and USD · API access · free 7-day trial, no card required.