Our official statistics agency published the first findings of its survey of earning and spending by SA households this week. These show a nation that is still highly unequal, on race and gender lines but also on education and geographic lines. But the more striking finding is that the gap has narrowed dramatically over the past two decades. And where race clearly is the divide, which Stats SA tends to focus on, along with gender, it’s clear from the survey that education is now a bigger source of inequality than race. Employment versus unemployment would no doubt have emerged as an even bigger one, had the statistics agency done that breakdown.
This is the first full income and expenditure survey (IES) Stats SA has conducted since 2010/11. It’s based on a survey of more than 19,000 households. The data is used to calculate everything from the weights in the inflation basket to official inequality and poverty ratios. In other words, it goes to the heart of SA’s challenge of how to raise living standards for all.
Household surveys are a challenge at the best of times, and Stats SA seems lately to find them more challenging than most. Big questions have been raised about the quality of its most recent census data. The IES raises some questions too. But imperfect though it may be, it captures trends and patterns that provide an important window on a changing society.
The survey shows households headed by those with tertiary education earned almost seven times those with no education and almost six times those with primary education.
The headline national findings are the most depressing. SA’s progress on raising living standards has been pretty pathetic since 2006, when the first IES was conducted. If you believe the numbers, household spending has declined almost 2% in real, inflation-adjusted terms over the period, and household incomes have risen by just 5%. That’s total, not per year. Why income and expenditure have gone in different directions is not clear; it seems implausible that households are saving as much as the survey implies.
What is clear is that while income inequality along racial lines remains stark, the racial mix has shifted dramatically. White-headed households had average incomes that were still 4.7 times those of black African-headed households in 2023. But where white average incomes have declined almost 8% in real terms since 2006, black African incomes have jumped by 46% over the period, and these households aren't the only group that has seen consistent real increases. Incomes have also jumped for coloured and Indian households. The consumer economy now looks quite different to what it was in 2006, with white households now accounting for just 25% of total consumer spending, down from 46%, while black African households’ share has climbed to 62%.
The gender gap has closed too, even though it remains wide, with the income of male-headed households declining by 1.5% in real terms over the past two decades while that of female-headed households has risen by 36%. But while race and gender tend to be much of the focus for Stats SA, the data on education is at least as interesting — and more useful for policymakers, who can’t change the race or gender people were born with, but can and should be doing far more to improve education outcomes given how strongly these are correlated with people’s life chances. The survey shows households headed by those with tertiary education earned almost seven times those with no education and almost six times those with primary education. There too, inequality has narrowed over the past two decades, with the incomes of the least educated households growing faster than the most educated.
Where it has widened is between the fastest and slowest growing provincial economies and cities, with the Western Cape, and Cape Town itself, pulling far ahead of other provinces and metros. And while Gauteng is still up there as the second richest province behind the Western Cape (together the two provinces account for over half of SA’s total consumer spending), it is sliding, as is Johannesburg.
Stats SA has yet to release a second report on the survey, which will provide its updated calculation of SA’s Gini coefficient, which measures inequality, and its poverty lines. Those will be closely watched, especially given that SA is reputed to have one of the world’s highest inequality levels, along with one of the world’s highest rates of unemployment (the two are so closely linked that it is disappointing that the IES didn’t add breakdowns by employment status).
But how seriously should we take those figures? Any questions about the quality of the IES will also reflect on the accuracy of those crucial ratios. Stats SA seems to have been struggling a bit with household surveys in recent years, whether because its fieldworkers struggle to gain access or because of a lack of money or training. Getting people to disclose their incomes may be a struggle, even for the best trained of fieldworkers; asking them to remember what they spent on which item last week even more so, though the survey tries to address this by getting households to fill in spending diaries. Even so, the numbers raise some concerns. If households are spending only 70% of their incomes on average, as the survey shows, why is SA’s household savings rate so low and indebtedness so high? The apparent disconnect between expenditure and income numbers is a concern. Others may emerge as economists sift through the numbers.
Such surveys should always come with a health warning. But even if the exact numbers are imperfect the patterns and trends are illuminating — and those in this week’s IES could hardly be more important, not just to inform policy but to inform our picture of ourselves as a country. Instead of being defensive, as Stats SA has been lately in response to criticism, it could simply be honest about the limitations of its surveys and help us interpret them accordingly.
• Joffe is editor-at-large.
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.