We treat the informal economy as a waiting room. For most of the planet, it is the main hall.
Start with the number that should reframe everything. Roughly two billion people, close to 58 percent of everyone on earth who works, work informally. Not at the margins of the economy. As the economy.1
In sub-Saharan Africa the figure climbs toward nine in ten workers.2 Yet our instincts, our policies and our pitch decks still treat the formal payslip as the default and everything else as a temporary condition people are waiting to graduate out of. The data says the opposite. Informality is not the absence of an economy. It is the presence of a different one.
South Africa sits inside this story with its own twist. Of the 17.1 million people Stats SA counted as employed at the end of 2025, the formal sector held 72 percent. The informal sector was the second largest source of work at 21.4 percent, with a further 6.6 percent in private households.3 Put plainly, more than one in four working South Africans earns a living in the part of the economy our headline statistics treat as a rounding note.
Reported On that 21.4 percent ›
And here is the first crack in the instrument: that 21.4 percent was calculated under a brand new definition adopted in 2025. Stats SA itself cautions that the figure cannot be compared with last year's.3 We did not just measure the informal economy. We changed the ruler mid-measurement.
The teal and gold together are the 28 percent of South African work that lives, by design, outside the formal record. Stats SA, QLFS Q4 2025
A statistic is a decision about what deserves to be counted. We decided wrong.
This brief is about what the economy produces and official measurement cannot see: the output, employment and loyalty that never reach a national account. It reads the flow of goods and labour.
Its companion, Africa's Hidden Capital Markets, maps the other half: how capital is raised, allocated and priced outside formal markets. One follows the flow of output, the other the flow of funds. Read together, they close the gap the dashboard leaves open.
Stack the economy by how clearly each layer shows up in official measurement. Then watch what happens to growth and loyalty as you go down.
This is the idea the rest of the brief rests on. Order the economy's layers from the most visible to measurement at the top, the formal payslip counted to the cent, down to the least visible at the bottom, the creator invoicing a foreign client from a bedroom. Now read two channels off each layer at once: how brightly it shows up on the national dashboard, and how much vitality, growth and loyalty, it actually carries. The two channels do not move together. They lean in opposite directions.
↓ Read the stack from the top down. The left column, how brightly each layer registers in measurement, thins as the right column, how fast it is growing and how much loyalty it holds, thickens. Where the two cross is the whole argument.
As a layer of the economy becomes harder to measure, it tends to become more important, not less. The brightest part of the dashboard sits over the most inert part of the market. The darkest part sits over the fastest growth and the deepest loyalty.
Visibility and vitality scores are a Diagelo composite, built from the measurement methods and sector signals cited throughout this brief: official coverage on one channel, reported growth and loyalty on the other. They are illustrative of direction and order, not precise magnitudes. The data they summarise is real and sourced below.
Informality is not one economy hiding in the dark. It is several, moving through instruments built for someone else.
Before mapping anything, one piece of housekeeping, because it is where most analysis quietly goes wrong. The word informal is carrying far too much weight. Underneath it sit several different things that happen to share a label, and the moment they are blurred into one number, the number lies. So here is what the term is actually hiding.
Informal employment
labourInformal sector
enterpriseTownship economy
placeStokvel flows
financeShadow economy
uncertaintyThese are not five neat tiers; they overlap, bleed into one another, and refuse to line up cleanly. That is rather the point. A figure that captures one of them gets quoted as if it captured all five, and a market gets miscounted in the gap between the definitions. Hold the distinctions loosely, but do hold them.
When people picture the informal economy they picture a street vendor. That image is a generation out of date.
The under-measured economy now runs across four very different terrains, each with its own logic, its own rails, and its own gap in the official record. Tap through them.
→ Open all four. The figure is the headline; the line beneath it, what the dashboard misses, is the one that costs money.
The newest and least-measured terrain is digital. The global gig economy was worth about US$556.7 billion in 2024 and is projected to more than triple. South African creators and freelancers plug into it directly, though the country still trails peers like Kenya and Nigeria in online gig work.
What the dashboard missesOfficial data sees almost nothing cleanly. It misses foreign-currency micro-exports, platform dependency, irregular but real income and new youth income pathways.
“A laptop, a fibre line and a foreign client. The whole transaction can happen without a single South African statistic ever noticing.”
Ask three credible institutions how big the informal economy is, and you will get three answers that disagree by a factor of six.
This is not incompetence, and it pays to be precise about why. National accounts, survey estimates and shadow-economy models are not aimed at the same object; each catches a different slice and is structurally blind to the rest. But there is a harder dynamic underneath, and it is the part most write-ups skate past. Measurement follows money. The parts of the economy that attract formal capital throw off the records that make them easier still to measure, while the parts that run on cash and trust stay dark precisely because nothing official is watching them. Legibility compounds. So the gap is a trap that deepens on its own, not a fixed blind spot you close by widening the lens. Drag the slider and watch the same country redraw itself.
The narrow national-accounts view sees the informal sector as a small slice of value added. World Economics' survey method lifts it to roughly a quarter of GDP.10 Run Schneider's MIMIC model across sub-Saharan Africa and the shadow economy averages closer to 38 percent.11 Same activity. Three rulers. The gap between them is not noise. It is the size and shape of our blind spot.
Most organisations do not ignore this economy. Their instruments are trained not to see it.
The blind spot is architecture, not carelessness. Corporate dashboards privilege the signals the formal system emits, card spend, payslips, registered merchants, and discount everything that moves on cash, trust and the group chat. What the system cannot read, it treats as if it does not exist.
When the instrument cannot see the right column, demand is underestimated, creditworthiness is mispriced, distribution is misdesigned and loyalty is misread. The result is not bad measurement. It is bad decisions.
You cannot fund, tax, serve or partner with an economy you have refused to measure.
The future economy already exists. We are simply not measuring it properly.
That sentence sounds soft. It is a balance-sheet problem. Every organisation building a South African plan on official figures alone is steering by an instrument that omits the part of the market with the deepest loyalty, the fastest growth, and the least competition for those who learn to read it.
The spaza-registration drive of late 2024 into 2025 is the tell, though I want to be honest about how much weight it can bear. Of roughly 50,000 applications, about 19,000 were approved; afterwards, 41 percent of township residents reportedly said they trusted spazas more.9 That last figure comes from a single drive readout, not a longitudinal study, and I would not stake a thesis on the exact number. What is harder to wave away is the direction. A market that grows more trusted as it is counted is not a market waiting to be formalised before it becomes real. It is real now, and trust, not registration, is its currency. The firms that win here will be the ones who measured it first.
What this means
An instrument that cannot see most of the economy is not a neutral problem. It hands an advantage to whoever fixes it first, and quietly taxes everyone who does not. The same finding reads differently depending on where you sit.
For investors
The deepest loyalty and the fastest growth on the continent sit in the layers your diligence deck cannot price.
Treat the measurement gap as an information edge, not a risk to be avoided. The logic is uncomfortable but clean: a market no model can see is also a market no competitor has cornered. Township commerce alone is widely put near R900 billion a year, with credible analysts reaching past a trillion, and informal retail is reportedly outgrowing the listed grocers it is supposed to sit beneath. None of that resolves cleanly on a screen built from listed comparables and bureau data, which is precisely why the entry multiples have not been bid up.
The work, then, is to build the instrument the market lacks. Underwrite the rotating-savings group as the credit institution it already is. Price demand from cash velocity and trading networks, not card spend. The firms that compound here will be the ones that measured the place before it was fashionable to, not the ones that arrive once the data has caught up and the margin has gone.
For founders
Build for income that is irregular but real, and for trust the bureau never wrote down.
The household that pools through a stokvel, repays on social trust and earns across three uneven channels is creditworthy in ways no bureau file will ever show. That is a product brief, not a hard-luck story. Design around how money actually moves, the WhatsApp order, the cash float, the month-end obligation to the group, not the rails the formal system happens to emit.
And resist the urge to formalise your customer before you serve them. The spaza-registration drive cuts both ways: registration has its uses, but the thing that earns custom here is trust, not a certificate. Build the trust into the product, and formalisation becomes a feature you can offer, not a gate you make people climb.
And, more briefly, for everyone else
You cannot tax, protect or plan for what you refuse to measure. Fund the instrument before the next formalisation drive: count the informal sector as a first-class economy, not a residual line.
The stokvel, the spaza account, the income invoiced in dollars are not the waiting room before the real economy. For most of us, they are it. Ask why instruments paid for with public money still cannot see the work that feeds your street.
The richest unsolved measurement problem on the continent is sitting in plain sight. The methods that finally read this economy will be exported, not imported.
Stop convening the formal economy and calling it the economy. Taxi ranks, trader associations and rotating-savings groups already move value. They are infrastructure. Build with them, not around them.