Migrant exodus exposes the hidden truth behind SA’s unemployment paradox

This winter, in KwaZulu-Natal, sugar cane is standing uncut because there is nobody to harvest it. This same winter, 8.1 million South Africans are looking for work and cannot find it. In one of the most unemployed societies on Earth, work is going undone due to what seems to be a labour shortage. At the onset, this looks like a paradox within the labour market. However, the truth is that it is not a labour market with a problem, but a labour market that has stopped being a market.South Africa has spent a decade arguing about how this could be true, at lower intensity, because the shortage was chronic rather than acute. The official unemployment rate is 32.7%, and 43.7%, once discouraged work-seekers are counted. And yet farmers have insisted, year after year, that they cannot find workers. For years the debate ran on assertion. Farmers said locals will not do the work and unions countered by saying that farmers will not pay for the work. Nobody could prove anything, because the two claims could never be tested against each other.Then the test ran itself.The test nobody designedAfter the March and March movement’s 30 June deadline for undocumented immigrants to leave the country, thousands of foreign workers fled farms, factories and households, driven out by protests, community raids and intensified enforcement. In KwaZulu-Natal’s sugar belt, growers estimate they lost around 30% of their cane cutters almost overnight. Reports describe citrus hanging unpicked mid-harvest, from the groves outside Durban to the Eastern Cape. This is the closest thing to a natural experiment our labour market will ever produce: remove the migrant workforce and see whether the unemployed fill the gap. Granted, it is a dirty experiment. The same shock that removed the workers also changed the enforcement climate, the safety calculus and the harvest timeline, so nothing was held constant. But it is the only experiment we will ever get.What the farmers seeThe farmers’ reading goes like this. Growers report they are not getting enough willing locals, and that the locals who do arrive cannot match the productivity of experienced crews, because cane cutting is semi-skilled work that depends on pace and technique. I will add my own weight to that from experience rather than theory. I’ve run harvest crews on vegetable fields in the Free State: cabbage, horseradish, pumpkin and squash. Cutting cabbage looks unskilled until you watch someone do it badly. There is a judgement in every head, a knife angle, a pace that has to hold from first light to knock-off, and a rhythm to loading that a good crew settles into like a machine and a green crew never finds in its first season. Horseradish will humble a strong man by mid-morning. A green crew is not a slightly slower version of an experienced crew. It is a different order of output entirely, and a crop in the field does not wait for anyone’s learning curve. On this reading, the vacancies were never really available to the unemployed, and the word “vacancy” hides more than it describes. A vacancy exists on a spreadsheet, where eight million jobless people and thousands of unfilled jobs sit in the same national column and look like a matching problem. Walk it down to ground level, and it dissolves. The job demands embodied skills that the unemployed man was never given a chance to learn. It is hours away from where apartheid planning put his home, in a rural district with no transport he can afford. And once you subtract the taxi fare and the cost of living near the farm, a six-day week of brutal work nets him barely more than staying where he is, close to his family, his network, and whatever the household already pools. Calling him unwilling is arithmetic, not attitude. He is not refusing work. He is refusing a loss.The floor under the refusalThe social grant system belongs in this arithmetic, and it deserves more honesty than either side gives it. The lazy version of the argument says grants keep people out of the fields. That version fails on the numbers: nobody funds idleness on the R370 a month the Social Relief of Distress grant pays, an amount frozen since 2024. The accurate version is about household floors. A pension of R2,400 and two child grants of R580 give a rural household a small, certain monthly floor, and a net-loss job competes against that certainty and loses. The grants are not making anyone lazy. They are making the true price of labour visible, by giving people just enough footing to refuse work that costs more than it pays. Step back far enough, and the grant system is something else as well: the state’s standing admission that the formal economy cannot absorb its people. South Africa expects to pay out some 26.5 million social grants a month in this fiscal year, most of them for children and pensioners, in an economy with 16.8 million jobs. Set that beside the farm labour story and the symmetry is uncomfortable. For 20 years, two subsidies propped up the same equilibrium from opposite ends: migrant precarity subsidised the cost of production, and social grants subsidised the survival of the workforce that production could not employ. Neither subsidy ever appeared on an invoice. However, both were ways for the country to avoid paying the true price of labour.What the unions seeThe unions’ reading is the mirror image. The Agricultural and Allied Workers Union argues there is no shortage of willing South Africans, only employers who prefer migrants because they are cheaper, more flexible, and less likely to demand contracts or legal protections, with foreign farmworkers routinely paid below the national minimum wage. The Casual Workers Advice Office has logged roughly 3,000 reports of migrant workers dismissed, intimidated or told not to return. On this reading, the experiment is rigged. What vanished was not labour; it was underpriced labour, and the “shortage” is simply the market refusing to pay the legal rate.Growers dispute the underpayment claim. Pratish Sharma, a North Coast grower and SA Canegrowers board member, has said publicly that foreign workers were employed like anyone else and were not paid any less. I cannot adjudicate wage records from here, and neither can the unions or the growers speaking in aggregates. But the underpayment question, important as it is, does not settle the deeper one. Even where every wage was legal, precarity pays in other currencies. The AFP reporting describes foreign farmworkers working seven-day weeks and skipping breaks to maximise piece-rate earnings. An undocumented worker cannot refuse, cannot complain, and cannot leave for a better offer without risk. That compliance has economic value whether or not the hourly rate is lawful.Where the skill premium wentAnd here the two readings collide in a way neither side wants to acknowledge. The farmers’ own testimony is that experienced cane cutters hold scarce, valuable skill. That is their entire explanation for why locals cannot simply replace them. But in any functioning labour market, scarce skill commands a premium, and these workers were earning minimum wage at best. So where did the skill premium go? It was captured further up the chain, because legal precarity destroyed the bargaining power that normally converts skill into pay. An undocumented expert cannot bargain. On piece rates, a skilled cutter did out-earn a green one, but only by cutting more tonnes, which is precisely why we read of seven-day weeks and skipped breaks. The premium existed; the worker funded it with his or her own body. Skill plus power earns a higher rate. Skill without power earns the privilege of working harder. The farmers are right that the skill is real. The unions are right that it was never compensated as skill. The mechanism that reconciles both is power, not the market.The wage that does not existMy own view is that both readings are true. The question worth asking is what follows when they are true at the same time.If the union reading is even partly right, then a meaningful share of South African agriculture’s cost base rested on labour that was cheaper than a formal local workforce, whether through unlawful wages, lawful wages with unlawful intensity, or simply the flexibility that fear provides. If the farmer reading is also right, then that same agriculture cannot fund a formal replacement workforce at lower productivity and survive, because the margin does not exist. North Coast cane growers are facing prices down 10-12% against input costs up around 30%. One farmer said that hiring additional labour at current margins might close his operation. So the wage that would clear this market, high enough to draw local workers into hard rural seasonal work, low enough for a squeezed grower to fund, may simply not exist. That is not a labour market failure. That is a labour market telling the truth for the first time in 20 years.Who pays the invoiceAyabonga Cawe, writing in Business Day, read this moment as exposing the contradiction of free trade: we accept the free movement of capital, goods and information while restricting the movement of labour. It is a fair point. But I think the sharper lens is not trade theory; it is power. Ask the political scientist’s question: who bears the adjustment cost, and who had the power to shift it?For two decades, the answer was migrant workers. Their legal precarity was not an unfortunate side effect of the arrangement. It was what held the arrangement up. It let growers survive a price-cost squeeze that formal wages could not have survived. It let the government avoid choosing between food-price inflation and rural unemployment. It let consumers eat cheaply. The cost of making all of that work was carried by people with no vote, no recourse, and now, it turns out, no protection. When a political movement decided to make them the target, they absorbed that cost too, by fleeing.The exodus does not resolve the paradox. It redistributes the cost of it. The question for the next six months is where the cost lands next. If it lands on growers, we get farm consolidations and, in sugar’s case, real risk to milling infrastructure that is already fragile. If it lands on consumers, we get food inflation into an election-sensitive economy. If it lands on workers, local or foreign, we get either a genuine wage adjustment or a quiet regularisation that restores the old bargain under a new name. My money, for what it is worth, is on the last option. Systems that have externalised a cost for 20 years rarely choose to internalise it voluntarily. They find a new way to hide it.The honest conclusion from the experiment, so far, is uncomfortable for everyone. Three weeks is early, and I will happily revise this if spring proves me wrong. But the unemployed have not surged into the fields, so the pure exploitation story is incomplete. And the farms cannot pay their way to a local workforce, so the pure reservation-wage story is incomplete too. What the experiment actually revealed is a sector whose economics were quietly subsidised by the legal vulnerability of the people working in it. That subsidy has been cancelled. Someone must now pay the invoice.We should at least be honest about who was paying it before.The cane is still standing. The queue is still 8.1 million people long. Everything this country refuses to price honestly lies in the distance between them. DM