In South Africa’s political language, migrants appear almost entirely on the cost side of the ledger: a burden on services, a pressure on wages, a number to be reduced. The economics point the other way. If the migrant contribution to national output is anywhere near the order of magnitude this debate implies — a share often discussed in the region of nine percent — then it is not a rounding error to be wished away. It is a slice of GDP that no finance ministry, anywhere, treats as disposable.
The gap between that contribution and the language used to describe it is the real subject. A cost narrative justifies removal; an output narrative demands scenario planning. The two framings are not neutral descriptions of the same facts; they point policy in opposite directions, and only one of them survives contact with the national accounts. The recent protests, and the economic blowback they risk, force the question of which narrative South Africa can actually afford.
The Ledger: Cost Narrative Against Output Reality
A contribution to GDP is not charity received; it is value produced. Migrants in South Africa work, and in working they generate output. They consume, and in consuming they create demand that sustains local jobs. They rent, and in renting they underwrite landlords and property income. They run microenterprises that employ South Africans, and they pay value-added tax on almost everything they buy. Each of these is a line in the national accounts, whether or not the political conversation books it there.
The contribution also compounds. A wage earned is largely a wage re-spent — on rent, transport, food and school fees inside South Africa — so each rand of migrant income circulates through local businesses before it settles. That multiplier is precisely what a cost narrative cannot see, because it counts the service a migrant uses but not the second and third transactions that same migrant sets in motion. The cost framing captures only part of the picture — the fiscal expense of services used — and omits the far larger production and demand contribution. It is the accounting equivalent of reading only the expense column of a business and declaring it unprofitable. When a share of output in the order of nine percent is described purely as a burden, the description is not measured analysis; it is a category error with a policy attached.
Treating production as if it were only cost is not a value judgement — it is a bookkeeping mistake.
The Scenarios: Modelling a 5%, 10% and 20% Contraction
The constructive response is not to assert a single number but to test a range. Operators, lenders and public planners should model revenue and employment under a 5%, 10% and 20% contraction of the migrant economy, and read the results before, not after, events force them. Three tiers capture three distinct futures: mild attrition that trims demand at the margin, serious disruption that removes labour and consumers together, and systemic contraction that reprices whole sectors.
The tiers matter because the effects are unlikely to be linear. A five percent contraction may be absorbed through thinner margins and modest belt-tightening; a twenty percent contraction can cross the threshold at which a supplier loses viable volume, a landlord faces a rental void that reprices a whole block, or a lender’s informal-sector book tips from stressed to impaired. Scenario work exists to locate those thresholds before they are met. The value of it is that it converts a charged political variable into a planning input. A retailer can ask what a ten percent fall in migrant consumer spending does to township turnover. A lender can ask what a twenty percent contraction does to a book weighted toward informal-sector borrowers. A provincial planner can ask what either does to VAT receipts and local employment. None of these questions require a precise contribution figure to be useful; they require only that the possibility be modelled rather than dismissed.
You do not need the exact number to model the risk; you need only the willingness to run the scenario.
The Fiscal Line: Who Actually Pays
The fiscal consequence is where the cost narrative most clearly inverts. Migrant enterprises employ South Africans, and their supply chains run through South African wholesalers, landlords and transporters. A contraction in that activity does not return neatly to the fiscus as a saving; it shows up as lost VAT, lost output and, in the sectors that cannot quickly rehire, additional unemployment against a rate that already reached 32.7% in the first quarter. The removal of a taxpaying, employing, consuming population is not a budget relief. It is a revenue event.
The timing makes it worse. Tax foregone lands immediately, while any offsetting benefit — a local worker eventually filling a vacated role — arrives slowly, if at all, in a labour market that has struggled to create jobs for a decade. A fiscus that trades a certain, near-term revenue loss for an uncertain, distant gain is not saving money; it is deferring a shortfall and calling it prudence. Government’s own cabinet communications acknowledge the economy’s fragility even as the political weather turns against the people partly sustaining it. That tension is not rhetorical. A treasury cannot simultaneously plan for growth and treat a measurable share of the productive base as expendable without the arithmetic eventually presenting the bill.
There is no version of the fiscal maths in which removing taxpayers and employers reads as a saving.
The Institutional Test: Language Against Arithmetic
The deeper test is institutional. Stats SA, the Reserve Bank and the Treasury deal in arithmetic; the political conversation deals in language. When the language treats an output share as disposable and the arithmetic treats it as load-bearing, the two cannot both govern policy. One of them is describing a country that does not exist.
For decision-makers, the resolution is to trust the arithmetic and instrument the language as a risk. Track the political framing as a driver of instability, but plan the business and the budget on the measured contribution — because that is the number the accounts will ultimately enforce.
When language and arithmetic disagree about a tenth of the economy, it is the arithmetic that sends the invoice.
So What: Run the Scenarios Before the Politics Runs You
The intelligence angle is specific and actionable. Build revenue and employment scenarios for a 5%, 10% and 20% migrant-economy contraction into planning now, whether you run a retailer, a lender, a farm or a provincial budget. Treat the results as a map of exposure, not as a political statement, and revisit them as conditions change.
Nine percent of GDP — or whatever the true figure proves to be — cannot be treated as disposable, because output does not become weightless simply because it is politically inconvenient. The measured position is to hold both truths at once: the hostility is real and harms real people, and the economics of removing that population are severe and quantifiable. Institutions that model the second will be better placed to argue for restraint on the first.
