South Africa - Getting a grip on the informal sector

21 July 2025

The outgoing Capitec CEO, Gerrie Fourie, generated quite a debate on measuring the informal sector and unemployment in South Africa.

Image: JP Landman - Political & Trend Analyst

I paid a visit to Joe de Beer, Deputy Director-General responsible for economic statistics at Stats SA. He was very generous and helpful and clearly has his feet firmly on the ground. I also rely on several 30-year overviews of the SA economy by the Bureau of Economic Research at Stellenbosch and work done by Prof Johan Fourie at Stellenbosch. Any mistakes are mine.

The basics

The size of a country’s economy is determined through a system of accounts, standardised globally by the United Nations and known as the SNA or System of National Accounts. The first guidelines were published in 1953. The SNA has been regularly updated since then, with the most recent update in 2025.

The SNA measures the economy through a sequence of accounts. These allow, among others, the estimation of GDP in various ways, using production, income, and expenditure approaches. The accounts capture the money flow and give us a bird's-eye view of the economy. We call that bird’s-eye view GDP or gross domestic product.

GDP is the total value added in the economy by all economic actors, whether they're individuals, households, companies, mines, banks, the government … the whole lot.

This is very important for a discussion on the informal economy because value added is a net number. It's the combined value of all goods and services minus the cost of inputs to produce them. The deduction is very important; otherwise, we double count.

Let’s apply all this.

Follow the street vendor

An informal pavement seller sells tomatoes and sweets on the pavement. She sells it for R2 000 and that is her earnings for the day. She is not registered in any way whatsoever. Many now assume that her R2 000 earnings are not captured in the national statistics. Not so quick.

The pavement seller bought tomatoes and sweets from suppliers. It can be from the fresh produce market, or a branch of Boxer or Makro, a nearby corner vegetable shop, or perhaps a neighbour who grows tomatoes. Let’s assume her purchases came to R1 000, then her value added to GDP was R2 000 minus R1 000. She also had other input costs. Let’s assume she spent R50 on transport to her trading spot and another R50 to employ a youngster to help carry her stuff and look after the goods. Her value added to GDP is now R2 000 minus R1 000 minus 2 x R50 = R900. Very different from her R2 000 sales for the day.

Her input costs of R1 100 are captured elsewhere in the GDP process. It's captured in the retail sales where she bought her supplies, the petrol sales to the taxi she took, and the wage she paid to the youngster helping her.

The bottom line: to just take the R2 000 turnover and allege it is not included in the formal GDP number is inaccurate. One must deduct her input costs. The same logic holds for the vetkoek seller, a spaza shop, a shebeen, a hairdresser, a backyard mechanic, a physiotherapist, an estate agent, and a doctor working from home.

What about the R900 difference between the R2 000 sales and the R1 100 input costs? Is that recorded anywhere? She uses the R900 to buy personal groceries and clothes and to pay fees … basically, her living expenses. Her purchases are sales made by another party. In economic parlance, her purchases constitute demand, which the expenditure account captures. Likewise, the doctor or tennis coach who only takes cash will likely spend that cash at some point. Their production (cash earned) is converted into purchases (expenditure), which are recorded.

So, a large portion of the R2 000 is captured in GDP data, be it in the production or expenditure accounts.

Informal sector

The outline above suggests that most transactions will get picked up somewhere in the SNA.

Most, but not every rand. The question is then: how much remains uncounted?

Stats SA assumes that about 8% of the economy is informal and not captured by official statistics. In Joe de Beer’s opinion, this may be too low. The number may be adjusted upwards when the next 5-year benchmarking and rebasing of the various sectors in the economy is done. IMF people are currently at Stats SA to help prepare for that change.

The 8% may not sound like much. But the economy measured R7.024 trillion in 2023 – 8% of this would be R561.9 billion. That's more than the entire mining sector that year (R444 billion). Think about it: all the platinum, gold, coal, copper, diamond and other mines are smaller than the assumed informal sector! A lot of hairdressing, street vending, cash visits to the doctor and so on can fit into that amount. It's a massive sector, which many companies have not explored. Capitec has, and very successfully so.

Unemployment

According to Stats SA, unemployment in the first quarter of 2025 came to 32.9% or 8.2 million people, while 16.7 million people were employed. These numbers got the Capitec boss animated.

The numbers are derived from a household survey that Stats SA conducts every quarter in line with international criteria. It covers 33 000 dwellings, which are surveyed every quarter. A criticism is that the selection of 33 000 is based on 2011 census data. That is true, but it has also been updated with community survey data from other Stats SA reports, and the sample is updated frequently. Stats SA is now developing a new sample, using information gleaned from the last census.

Capitec has 24 million individual clients, of whom 8.8 million are fully banked, and 13 million are active app users. Put that in a country context: 16.7 million people employed; about 18 million on social grants; and about 9 million on the Covid-19 relief grant. That's close to 44 million people. Almost double the number of Capitec clients.

Clearly, Capitec is sitting on a treasure trove of information (as do other banks). But this on its own cannot tell us who are employed, unemployed, grant beneficiaries, and so on. More work needs to be done to determine this. Let’s see if the Fourie debate can inspire alternative methods of measurement, other than household surveys. If successful, it would be a win-win.

So What?

• In measuring an entire economy, one will never have 100% accurate numbers. The trick is to capture most of it.
• Certainly, anecdotal evidence from one's own experience is not enough to draw firm conclusions on something as big, diverse and complicated as an economy with more than 62 million people. One needs to be on firmer ground.
• The system of national accounts run by Stats SA is robust, complies with global standards, and is regularly updated.
• The current inclusion of the informal sector in national GDP numbers is 8%, bigger than the mining sector. The informal sector is certainly not ignored in the statistics, even if some businesses do.
• More rigorous work on unemployment will be very useful. Banks can usefully collaborate with Stats SA on this.

Sources

1. Interview, Joe de Beer, Deputy Director-General: Economic Statistics, Stats SA, July 2025.
2. Thirty years of democracy, Bureau for Economic Research (BER), August 2024.
3. Fourie v Stats SA: A final word, Prof Johan Fourie, Currency, July 2025.

Coutesy: JP Landman