With South African municipal elections set to take place in 2021, we can expect a lot of service delivery comparisons. They may not all add up.
Did you switch on a light today? Pour yourself a glass of water? Flush the toilet? Access to electricity, water and sanitation are often key measures when assessing progress in a country.
In South Africa there is a common year of reference: 1994. It marked the country's first democratic election. Politicians have a habit of measuring progress - or lack thereof - from this date.
Over the years, Africa Check has fact-checked numerous claims about access to services between 1994 and the present day. Recently an Africa Check reader asked us to fact-check the claim that fewer people in South Africa have access to "clean water" in 2020 than in 1994.
To check this you need access to reliable data. But it is harder to come by than you might think.
Data from early 1990s limited in number and use
"Before 1994, most surveys, including the census, did not include Africans in full and did not weight samples for them appropriately," Neva Makgetla, senior economist at the Trade and Industrial Policy Strategies research institute in Pretoria, told Africa Check.
Surveys also excluded the so-called independent homelands - areas designated for particular ethnic groups - where a large share of the African population lived. And because of the trade and financial sanctions imposed by other countries in response to apartheid, economic data was unreliable, Makgetla said. Without a full census, private surveys also faced issues.
"It took a while before Statistics South Africa could develop systems to fully count the African population and develop better weights for surveys," Makgetla said. She added that many questions asked in these surveys were poorly designed.
"As a result, it is virtually impossible to compare demographic and socio-economic data from 1994 with any later findings."
Access to water in 1994 vs 2018
Let's return to the reader request to fact-check the claim that the number of people who had access to clean water was lower in 2020 than in 1994. It seemed straight-forward enough. We just needed to find the relevant figures for 1994 and 2018, the latest year for which data is available.
The most recent figures are easy to find. According to Statistics South Africa's, or Stats SA's, 2018 general household survey, 89% of households had access to piped or tap water. This included access in their dwellings, on site or off site.
For comparison you might reach for Stats SA's October 1995 Household Survey, which found that 78.5% of households had access to "clean water". But Niël Roux, the agency's manager of service delivery statistics, has previously explained that the survey was "hamstrung by a series of methodological and practical issues". The weighting issue, while less severe in the 1996 census, was a huge problem with the October Household Survey. This makes it difficult to compare its results with more recent statistics, Makgetla said.
A smaller nationally representative survey, carried out by the Southern Africa Labour Development Research Unit (SALDRU) in late 1993 and early 1994, found that 76.4% of households had access to piped water. This referred to municipal water and its quality would vary with location, Lynn Woolfrey, manager of the DataFirst research unit at the University of Cape Town, told Africa Check.
The available data makes it difficult to draw any clear conclusions about changes in access to clean or piped water since 1994. The situation is only slightly clearer when it comes to electricity.
Understating access in 1994
In 2019, president Cyril Ramaphosa claimed that only 36% of the population had access to electricity in 1994 and that the figure had since grown to 80%. He was wrong.
Access to electricity is measured in a number of different ways. It could involve counting how many houses are connected to the grid. But it could, and some have argued that it should, also include how many households are able to afford electricity and how many are able to use the full range of electricity services, for example lighting, cooking and heating.
According to Stats SA, their "earliest credible data" on electricity was from the 1996 census, given the problems with the agency's October 1995 survey highlighted above.
The census found that in 1996 58% of households used electricity for lighting. Almost 48% used electricity for cooking while 46.5% used it for heating. The 1993 survey by SALDRU estimated that 54% of households had access to electricity from the grid.
Both data sets indicate access to electricity in 1994 sat at above 50%. By understating access in 1994, Ramaphosa overstated the government's progress in increasing access to electricity.
Accuracy in public debate
With South African municipal elections set to take place in 2021, we can expect more of these kinds of comparisons. When campaigning, politicians need to be aware of the available data sources and their limitations.
We've put together this factsheet, detailing everything we do (and don't) know about service delivery in 1994, to ensure as clear a picture as possible. It is also a resource for voters who wish to fact-check election claims.