Stats SA’s latest general household survey (GHS) found that 87.4% of households had access to piped or tap water in 2025. This includes taps in the home or yard, neighbours’ taps and public or communal taps.

Access varied widely across the provinces. It was highest in the Western Cape (99.5%) and Gauteng (98.3%), and lowest in the Eastern Cape (67.7%) and Limpopo (63.2%).

Access has also fallen in some provinces since 2002, when Stats SA began collecting this data. These include Limpopo, Mpumalanga, the Free State and Gauteng. 

But access does not necessarily mean reliable supply. 

According to the GHS, 56.8% of households experienced water interruptions in 2025. Interruptions lasting at least two days were most common in the North West (61.5%), the Northern Cape (56.1%) and Mpumalanga (55.7%), and least common in the Western Cape (6.2%) and Gauteng (32.5%).

The ANC’s claim is correct, but its figure says more about connection to water infrastructure than about whether water is consistently available. 

Stats SA defines improved sanitation as having a flush toilet connected to a public sewerage system or septic tank, or a pit toilet with a ventilation pipe.

In 2025, 84% of households had access to one of these forms of sanitation, supporting the ANC’s claim.

Access was highest in the Western Cape (95.9%) and Gauteng (91.3%), and lowest in Mpumalanga and Limpopo, both at 64.8%.

Two thirds of households (67.6%) used flush toilets, while 16.8% used pit toilets with a ventilation pipe. But a further 13% still used pit toilets without ventilation.

As we’ve previously explained, the government uses “housing opportunities” as a broad term. It includes newly built state-subsidised houses, known as “breaking new ground” or BNG houses, and “serviced sites”, which are empty plots connected to utilities. 

The Department of Human Settlements (DHS) reports on housing delivery each year, but its figures do not always match. For example, in the 2023/24 annual report, DHS accounting officer Alec Moemi said the department delivered “42 233 BNG units and 19 859 serviced sites”. 

The same report later gives different figures: 20 223 serviced sites and 41 712 housing units. Yet another section repeats Moemi’s figures, while a table in an annexure lists 35 011 BNG houses and 21 952 serviced stands delivered across all provinces that year. It isn’t clear which figures should be used.

ANC: ‘Public sources do not clearly confirm the specific 451 000 figure’

Africa Check asked the DHS several times for an official tally of housing opportunities delivered, but had not received a response at the time of publication.

Across its annual reports for 2020/21 to 2024/25, the DHS reported delivering 185,484 BNG houses and 172 288 serviced sites. Together, these add up to 357 772 housing opportunities.

This is below the ANC’s claim of 451 000. The ANC told Africa Check that publicly available sources do not clearly confirm the 451 000 figure, but said: “it is directionally consistent with the annual delivery pattern, as in the Annual Performance Plans and Reports of DHS”.

The response pointed to DHS annual performance plans for 2025/26 onwards. But these set housing delivery targets, which are often not met. In 2024/25, for example, the department delivered 35 457 of a planned 38 069 BNG houses.

People who receive government-subsidised housing are meant to get a title deed - the legal document proving they own the property. 

In 1994, the government already owed title deeds to people who had been granted ownership of homes or land. As more housing has been built, it has failed to both clear this backlog and keep up with new title deeds.

The ANC told Africa Check that the manifesto figure was “plausible” but should be sourced from the DHS or National Treasury. It did not provide any title deed figures. 

The DHS also did not provide Africa Check with data on the number of title deeds handed out since 1994. 

Title deed delivery isn’t reducing the backlog

In October 2025, the human settlements minister told parliament that more than 1.2-million title deeds were still outstanding as of September 2025. This included 31 000 from before 1994. 

The backlog also appears to have grown over the past decade: in 2016, the department put it at just over 800 000.

In its 2014 election manifesto, the ANC promised to clear the title deeds backlog. It had not done so by 2019, when it repeated the promise. The government’s Medium Term Strategic Framework, which set out national goals for 2019 to 2024, aimed to deliver 1 193 222 title deeds. That target was about as many as the ANC said were delivered in 31 years.

As the ANC told Africa Check, the claim seems reasonable, but remains unproven. 

Stats SA’s latest data found that the share of households that received a government housing subsidy rose from 5.6% in 2002 to 12.9% in 2025. The ANC’s claim is therefore correct.

This represents an average annual increase of 0.32 percentage points.

The survey also asked about the quality of state-subsidised housing. Nationally, 9.7% of households in this housing described their walls as “weak or very weak”, while 9.6% said the same about their roofs. These complaints were most common in KwaZulu-Natal, where 14.7% reported weak walls and 15.6% reported weak roofs. 

According to the GHS, 85.6% of households had internet access in 2025, up from 28% in 2010. On this score, the ANC’s claim is accurate. 

But “access” includes using the internet anywhere, such as at work, public wifi, an internet cafe, school or through a mobile device.

In 2025, only 20.6% of households had access at home, up from 10.6% in 2010. Much of this increase occurred from 2021 onwards. 

According to Stats SA, this gap suggests most households access the internet through mobile devices rather than fixed connections at home.

The Presidential Employment Stimulus (PES) was launched in October 2020, during the Covid-19 pandemic, to provide temporary work and jobs for unemployed youth. It funds public and social employment, work experience, and support for livelihoods.

By February 2026, the presidency reported 2.52-million “jobs and livelihood opportunities”. More than half were in basic education. 

As we have previously explained, these are mostly short-term opportunities rather than permanent jobs. Programme head Dr Kate Philip said they last just a few months. They include temporary work, help to start and support small businesses, and other initiatives. 

Amid South Africa’s unemployment crisis, the short duration of most opportunities has drawn criticism. The latest PES progress report also acknowledges that these programmes should not simply be seen as “bridges” into long-term employment, because that assumes lasting jobs are available afterwards.

In a 2026 interview, Philip said the programme would like to offer young people at least a year of work, but lacked the budget.  

Despite high youth unemployment, the programme’s budget fell from R12.6-billion in 2021 to less than a quarter of that in 2025.

Stats SA’s latest employment figures cover April to June 2026, the second quarter of the year.

They show an unemployment rate of 62.8% among young people aged 15 to 24. This means they had no work but were available and actively looking when surveyed. 

The rate is higher when other groups of jobless young people, which Stats SA has begun reporting on, are included. The agency’s composite measure of labour underutilisation counts unemployed people as well as: 

• people who worked fewer hours than they needed
• people without work who were either available to work but not looking, or looking for work but not available to start

This gives a fuller picture of the shortage of work, and in South Africa, almost three quarters (73.8%) of young people aged 15 to 24 fall into this category. 

Stats SA also reports on people who are not in employment, education or training (neet), which was 45.4% among those aged 15 to 34. This has changed little over time: the figure was 45% in the second quarter of 2022.

The ANC’s claim is correct, even if it refers to two different age groups.

The South African Police Service publishes crime statistics every quarter. Researchers look at both the total number of murders and the murder rate - the number of murders per 100 000 people. This allows them to account for the country’s population size. 

The murder rate was 8.5 per 100 000 people in April to June 2026, the first quarter of the 2026/27 financial year. The ANC launched its manifesto in August, a few days before the crime data was published.

In the three quarters before that, it moved from 9.2 to 10.0, then 8.2. Strictly, these changes do not reflect a decline “across three consecutive reporting periods”. 

But researchers usually compare the same quarter across different years, because crime patterns vary throughout the year. For example, more murders are often recorded over the December holidays. 

Comparing April to June across years shows a steady decline in the murder rate: 10.1 in 2023/24, 9.2 in 2024/25, and most recently to 8.5. 

Experts cited by the Institute for Security Studies generally agree that murders are declining, but differ on why. Gareth Newham, special adviser to the acting police minister, has suggested this shows early success from policing that targets specific problems using evidence. Another view is that murder levels are returning to where they were before a surge after the Covid-19 lockdown and the July 2021 riots.

Separately, Jean Redpath, a researcher at the Dullah Omar Institute, has linked the decline to fewer power cuts.