Most people don’t pay much attention to the census. It sounds deathly dull and if they’ve ever encountered an enumerator, the questionnaire is tediously long and peppered with surprisingly invasive questions. But it matters enormously.
The snapshot of the population on census night determines government budget allocations, drives decisions about where to build schools and hospitals, and is used to gauge the impact of key programmes for years to come.
Given that so much hinges on the data, and that the 31% undercount in Stats SA’s 2022 census makes SA an international outlier, the criticism levelled at it in a report published last week by the Medical Research Council (MRC) must be taken seriously. The report was written by two of SA’s leading demographers, who say there are so many anomalies in Census 2022 that it is not fit for purpose.
UCT demographers Tom Moultrie and Rob Dorrington conclude the 62-million population estimate is 1-million too high and highlight peculiarities in the national and provincial figures, which are at odds with previous censuses and vital registration data such as that for births and deaths.
Crucially, they warn the census cannot be relied upon for the provincial equitable share formula, which is used by Treasury to determine budget allocations for provinces, districts and municipalities.
Stats SA has vigorously defended its work, saying the census was reviewed by a panel of experts, and it is thus confident in its results
To highlight the implausibility of the results, they provide a case study on Beaufort West, which supposedly saw its population surge 47.2% in a decade. But this isn’t borne out by satellite imagery of the land area covered by dwellings, which has barely changed. Nor is the apparent 54.9% surge in the number of people aged 18 and above supported by data from the voters roll, which saw registered voters increase just 12.9% since 2011.
In a country such as SA, where trust in the government is low and anxiety about personal security runs high, getting people to participate in the census is always a challenge. Stats SA conducted the 2022 census in even more difficult circumstances than usual: it took place in the midst of Covid-19, and in the wake of riots, local government elections and xenophobic violence fuelled by Operation Dudula.
But these logistical factors alone don’t explain the staggering undercount — which was more than double that of Census 2011 and the highest yet recorded by the UN Population Division. The demographers say Stats SA’s post-enumeration survey, which attempted to match household membership months after the census, was too small and ran too late to reliably calibrate the data.
Stats SA has vigorously defended its work, saying the census was reviewed by a panel of experts, and it is thus confident in its results. It says it followed international best practice and collaborated with other statistics offices running censuses in the same period.
But it hasn’t published its expert review or issued a detailed technical rebuttal to the concerns raised by the MRC report. It is thus not clear if their positions can be reconciled.
This leaves the government, and the Treasury in particular, in a pickle. It sits with a crucial planning tool that well-respected scientists say is unusable. There is only one thing to do: get the disputing parties round the table to fix the census or come up with a credible alternative. Too much money, and too many vital decisions, rest on it for questions about its validity to be allowed to linger.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.