The protein system
South Africa supplies 79.2 grams of protein per person per day, of which 36.5 grams, or 46 per cent, is animal-source.[1] Maize is both staple and feed grain, with 16.6 million tonnes produced in MY2025/26 against 14.2 million tonnes of consumption split almost evenly between food and feed, and 2.3 million tonnes exported.[2] Chicken is the animal protein, at 1.68 million tonnes forecast for 2026 against 1.92 million tonnes of consumption and 60 per cent of all meat eaten.[4]
The poultry complex is the largest in Sub-Saharan Africa and unusually consolidated. Worth R65.77 billion and 15 per cent of agricultural production, it has Rainbow and Astral Foods alone taking roughly half of broiler output, the top five producers supplying most of the feed used across the chain, and a broiler feed conversion ratio of 1.45 inside a national feed industry of about 13 million tonnes.[4][6]
Gross protein imports are 18.2 per cent of domestic supply protein and net of exports the balance is negative at -7.1 per cent, against record agricultural exports of USD 15.1 billion in 2025, up 10 per cent, on an agricultural trade surplus of USD 7.3 billion.[1][29] The system's failure is distributional. In 2023, a year of good harvests, 19.7 per cent of households experienced moderate or severe food insecurity on the FAO's Food Insecurity Experience Scale and 8.0 per cent severe food insecurity, up from 15.8 and 6.4 per cent in 2019, and press reporting of the 2024 General Household Survey put roughly one household in four in hunger.[7][30] Protein exists in this economy; about one household in five cannot reliably buy it.
Durable domestic capacity: 68
Quantitative score 73, adjusted down 5 points
The self-sufficiency arithmetic, computed by applying each commodity's protein content to FAOSTAT's 2023 production and domestic supply figures, is unusually strong for a middle-income country: 106.9 per cent on total protein, 93.2 per cent on animal protein, 99.8 per cent on eggs and 88 per cent on milk, with fish a net export sector at 493,000 tonnes produced against 386,000 tonnes of domestic supply.[1] South Africa is not alone in the region on the headline ratio, since the same computation puts Seychelles, Zambia, Tanzania and Malawi above 100 per cent, but it is much the largest surplus system and the only one with a heavily animal-weighted diet.[1] The gaps are wheat, capped by Western Cape rainfall at roughly half of a 3.86 million tonne market, and rice, which cannot be grown at scale at all.[2] Of 77 assessed marine stocks, 51, or 66 per cent, are not of concern, up from 20 of 43 in 2012, though that category includes 18 stocks whose status is unknown.[8]
We apply a downward adjustment of five points because active animal disease is degrading nominal capacity faster than the production series records. Foot-and-mouth disease has circulated since 2021 and by early 2026 had reached eight of nine provinces, with 207 outbreaks in KwaZulu-Natal of which 187 remained unresolved, spreading out of communal dip tanks into commercial beef and dairy herds and putting an R80 billion livestock sector at risk.[5] It reached Karan Beef's Heidelberg feedlot, the largest in the world and a system holding about 150,000 cattle at any time, in June 2025, and the Department of Agriculture concedes that resolution in KwaZulu-Natal is unlikely because buffalo in nearby reserves are carriers.[26][5] Avian influenza culled nearly 10 million birds in 2023, and the H5 vaccination protocol announced on 5 June 2025 had produced no vaccinated commercial flocks three months later, producers having judged the rules too costly to implement and Astral Foods being the only company then attempting to comply.[3][4]
Resource headroom: 58
Quantitative score 54, adjusted up 4 points
Water is the ceiling and it is close, and the land base underneath it is shrinking rather than static. Mean annual rainfall of 465 millimetres is half the world average, total annual runoff is roughly 49,000 million cubic metres, and more than 60 per cent of river flow comes from 20 per cent of the land area.[14] Renewable internal freshwater is 718 cubic metres per person per year, below the conventional 1,000 cubic metre scarcity line, and water stress has climbed every year from 62.9 per cent in 2018 to 67.6 per cent in 2022.[11][10] Irrigation takes 61 per cent of national water across 1.6 million hectares and it leaks, with canal conveyance losses of 30 to 45 per cent, non-revenue water up from 41 per cent in 2016 to 45.1 per cent in 2020/21, and a projected 17 per cent water gap by 2030.[14] FAOSTAT's own land-use series, last surveyed in 2011 and held flat since, records cropland falling from 14.20 million hectares in 2000 to 12.41 million and arable land from 13.81 million to 12.00 million; set against population growth of 34.0 per cent over the same period, that is a decline of 34.8 and 35.2 per cent respectively in per-capita terms, and arable land per person at 0.190 hectares is now slightly above the world figure of 0.172 only because the population comparison is favourable, not because the hectare count is growing.[12][45][46] There is no extensive land margin to open. Equipment tells a more mixed story: the area equipped for irrigation rose from 1.50 million hectares (10.6 per cent of cropland) in 2000 to 2.24 million (18.0 per cent) by 2021, but South Africa's own Department of Water and Sanitation puts area actually under irrigation at only 1.6 million hectares in 2023, so a material share of nominal capacity is going unwatered, which fits the water-stress trend rather than contradicting it.[45][14]
Three things earn an upward adjustment of four points, and none of them is a land or water expansion. Growth has come from yield, with maize area static at about three million hectares for seven years while output rose on seed genetics, agrochemicals and precision farming.[2] The Crop Estimates Committee put the 2026 commercial crop at 17.064 million tonnes in May 2026 and called it the largest on record, above USDA's March 2026 estimate of 16.0 million tonnes commercial for the same harvest.[21][2] The feed-protein base has been expanded by funded, operating capacity, with soybean area at 1.213 million hectares, a record 3.3 million tonne oilseed crush against processing capacity of about 3.7 million tonnes, and 2.28 million tonnes of meal produced against 1.88 million tonnes of domestic demand.[3] Fishmeal supplies about 0.8 per cent of protein meal demand, so the wild-fishmeal constraint that caps aquaculture-led systems barely applies.[3] And the capital behind that intensification is real: FAOSTAT publishes no agriculture-specific credit series for South Africa at all, only an undifferentiated Total Credit line, but the national accounts show total farming debt reaching R221.8 billion in 2024, up 8.1 per cent on 2023, against a capital asset base of R736.4 billion growing faster still, with 61 per cent of that debt held by the four largest commercial banks.[44][47] That capital funds intensification on land already in production, not a land-reform-driven expansion of the cultivated area: the Land Bank, the specialist lender to emerging and land-reform farmers, defaulted on R45.2 billion of debt in April 2020 and had cut its loan book to R17 billion by June 2024, of which over R8 billion is non-performing, and its Blended Finance Scheme for black commercial farmers, a ten-year programme with a minimum R3.2 billion government commitment, had approved R1.575 billion across 205 applications by July 2024.[48][49] Working for Water, the government's invasive alien plant clearing programme running since 1995, is a genuine mechanism for recovering water rather than land, having cleared over a million hectares cumulatively against invasive plants estimated to consume 1.44 to 2.44 billion cubic metres of water a year and cut national streamflow by nearly 3 per cent; the Western Cape alone has cleared 46,000 hectares since 2023, recovering about 15 billion litres a year for Cape Town, but national infestation is reported to be spreading faster than the programme clears it.[50]
Import exposure: 71
Quantitative score 76, adjusted down 5 points
Geography matters as much as production. No South African food import transits a maritime chokepoint. Its eight commercial seaports sit on open-ocean routes and handled about 304 million tonnes in the 2025/26 financial year, up 4.2 per cent on 8,630 vessel arrivals, the strongest growth since FY2011/12.[19] The port system that was the country's most-cited logistics failure has begun to recover: Durban gained 479 points in the 2025 Container Port Performance Index, from -721 to -242, the largest year-on-year improvement of any port assessed.[20] The base was very low and remains so, with Durban still 398th and Cape Town last.[20] Diversification on the largest dependency is wide: wheat imports of 1.842 million tonnes in MY2024/25 spread across ten origins led by Australia at 567,000 tonnes, Russia at 446,000 and Lithuania at 194,000, so no supplier exceeded 31 per cent.[2]
We apply a downward adjustment of five points because one concentrated exposure lands squarely on the poorest consumers. Brazil supplied 86 per cent of chicken meat imports in the first half of 2025, and mechanically deboned meat, the raw material for cheap processed protein, was 63 per cent of poultry imports in 2024, comes almost entirely from Brazil, enters duty free, and is not made domestically in volume because local processors prefer to sell the same carcasses as soup packs.[4] When South Africa suspended imports from the whole of Brazil on 16 May 2025 over a single-state avian influenza outbreak, partially lifting on 19 June and fully on 4 July, first-half imports fell 25 per cent year on year and individually quick frozen chicken prices rose 13.9 per cent.[4] Rice is similarly concentrated, with Thailand at 83 per cent, and the aggregate ratio conceals both.[2]
Upstream dependence: 64
Quantitative score 68, adjusted down 4 points
The most important upstream input has been substituted away. South Africa has moved from importing a large share of its protein meal to exporting a forecast 370,000 tonnes in MY2026/27 against 30,000 tonnes of imports, with 1.85 million tonnes of domestic soybean meal against 1.48 million tonnes of feed use, and USDA describes the feed industry as generally self-sufficient with little import of feed products or raw maize and soybean.[3][6] Energy, which troubled this system for a decade, is fixed for now: load shedding ended on 16 May 2025 and had not recurred for 441 consecutive days by early August 2026, with the energy availability factor at 66.97 per cent year to date and unplanned outages down 5,088 megawatts.[17] Cold chains and irrigation pumps are no longer rationed.
Residual dependencies are real but narrow. Potash is imported in full, with zero domestic production against 234,077 tonnes of K2O imports in 2023, while phosphate is the reverse case, 308,000 tonnes of domestic P2O5 covering 74 per cent of a 418,290 tonne requirement.[9] Nitrogen sits between, with domestic downstream manufacturing including Sasol's 400,000 tonne limestone ammonium nitrate plant, commissioned in 2012, and Omnia's nitric acid plant rated at 1,000 tonnes a day, against fertiliser imports of USD 824 million in 2023, of which nitrogen products were 56 per cent by value, drawn mainly from Saudi Arabia, Russia and Qatar.[31][32][22] Broiler genetics arrive as imported grandparent stock.[4]
We apply a downward adjustment of four points for veterinary inputs, because that failure is live and costing production now. The foot-and-mouth response depends almost wholly on imported vaccine: one million doses a month contracted from the Botswana Vaccine Institute in 2025, and by 22 April 2026 about nine million doses received in total, 2.5 million from Argentina and 3.5 million from Turkiye, with five million more on order.[5][23][24][33] Domestic production restarted at the Agricultural Research Council's Onderstepoort facility after a twenty-year gap, with a first batch of 12,900 doses handed over on 6 February 2026 against a stated ramp to 20,000 doses a week from March 2026 and 200,000 a week from 2027.[34]
Access and affordability: 38
Quantitative score 38, no adjustment applied
Access and affordability is South Africa's weakest pillar, an outcome the food-balance surplus does not on its own solve. The FAOSTAT/World Bank CoAHD unaffordability share stood at 57.9 per cent in 2025, little changed from 57.3 per cent in 2022 and down from a 2017 peak of 60.9 per cent.[36] Positioning within that band is pulled toward its lower half by three of the four supporting indicators. Prevalence of undernourishment has nearly tripled since 2001, from 3.6 to 10.0 per cent in 2023.[37] Child stunting has risen from an 18-year low of 22.1 per cent in 2018 to 24.4 per cent in 2024, a rate the WHO classifies as high severity.[38] Food price inflation has swung from 13.6 per cent in February 2023, the highest reading since April 2009, to 1.6 per cent in June 2026, an amplitude few peer economies match, with meat inflation alone falling from a 13.5 per cent peak in January 2026 to 5.1 per cent by June.[39][40]
The fifth component, physical access, does nothing to offset the other four. The apartheid-era spatial legacy leaves former-homeland and informal-settlement households at considerable distance from the formal retail that serves higher-income areas, and the supermarkets that do trade in low-income areas stock less healthy ranges than those in wealthier neighbourhoods.[41] Localised food insecurity reaches 45 per cent of households in Cape Town's Gugulethu and 36 per cent in Khayelitsha, and Statistics South Africa's 2024 General Household Survey recorded 22.2 per cent of households with inadequate or severely inadequate access to food, close to 14 million people.[41][30] The score therefore sits where the unaffordability share places it, and no higher. We apply no adjustment; the distributional evidence points the same way as the headline figure.
Shock endurance: 61
Quantitative score 58, adjusted up 3 points
South Africa holds no strategic reserve of grain. USDA records that the government maintains no maize, wheat or rice reserves and imposes no minimum stockholding requirement on anyone.[2] What exists is commercial: 2.255 million tonnes of maize and 606,000 tonnes of wheat, each about two months of use, and four weeks of rice, inside unmandated storage capacity above 20 million tonnes.[2] The 2016 drought defines the downside of holding no buffer: maize output fell to 7.78 million tonnes and imports reached 4.33 million tonnes in one year, against 436,000 tonnes in 2023.[1]
Crisis purchasing power scores 48. Gross international reserves stood at about 99 per cent of the IMF's own adequacy benchmark at end-2025 and are projected to cover around five months of imports, above the conventional three-month threshold.[42] Both Moody's and S&P have moved South Africa's outlook to positive, Moody's for the first time since 2007 and S&P via a one-notch upgrade to BB in November 2025, improving the terms on which the state could borrow to defend supply through a shock.[43] Gross loan debt nonetheless stabilises at a high 78.9 per cent of GDP in 2025/26 on gross national income of just USD 6,270 a head, and the Ba2/BB ratings remain two notches below investment grade, so South Africa's capacity to outbid wealthier importers on world markets is real but shallow rather than deep.[15][13][16][43]
We apply an upward adjustment of three points on documented shock performance. The El Nino drought cut the harvest 21.8 per cent to 12.85 million tonnes in 2024 and the system absorbed it on 794,000 tonnes of trade-year imports before recovering to 17.27 million, while poultry output was worked back from 19 million birds a week in 2023 to 22.6 million by July 2024 and freight rail from 149.5 million tonnes in FY2022/23 to an expected 168 million in FY2025/26.[35][27][2][4][18] R292.8 billion of social assistance in 2026/27, reaching 26.5 million beneficiaries and 12.6 million children on the child support grant, is the mechanism that actually protects protein access in a surplus economy.[15][28] It is also the weakest part of the case: 8.2 million of those beneficiaries hold the social relief of distress grant, whose allocation falls from R36.9 billion in 2026/27 to R1.7 billion in 2027/28, taking total coverage down to 18.1 million on Treasury's own medium-term numbers.[15] Consumers showed substitution capacity through the 2021 to 2024 squeeze, moving from individually quick frozen portions to offal and mechanically deboned cold meats as per-capita chicken consumption fell 8 per cent, which is adaptation and also evidence of stress.[4] Offsetting it is the uncontained foot-and-mouth outbreak, which cut beef exports 22.94 per cent in 2025 and Chinese sales by 69.2 per cent.[5][25] The indicators, combined with the physical evidence, put this pillar at 58: sub-investment-grade market access is the drag on an otherwise adequate reserve position, and a country with no physical grain buffer endures a shock on improving fiscal credibility rather than on the reserve depth of a currency-issuing economy.
Trajectory
Three of the four levers that matter here are already moving: electricity, subject to Eskom holding an availability factor in the mid-sixties;[17] ports and rail, recovering from a very low base;[20][18] and crushing capacity, which has turned feed protein from an import into an export.[3] The fourth is static: water is over-allocated, losses are rising, and the projected 2030 gap has an implementation plan without a funding settlement.[14] Nothing in current policy addresses the arithmetic that maize output swings by a fifth between a good year and an El Nino year on three million rainfed hectares.[35]
Two choices would improve this picture materially. The first is a strategic grain reserve, or a statutory minimum stockholding obligation on the private trade, sized against a repeat of 2016; more than twenty million tonnes of storage already exists, so the cost is working capital rather than construction.[2] The second is domestic veterinary vaccine manufacture at commercial scale, because foot-and-mouth and avian influenza are now the largest single drag on animal protein capacity.[5][33] One choice would worsen it: raising the poultry tariff wall further, already 62 per cent on bone-in cuts with anti-dumping duties on Brazilian product reaching 265.1 per cent, would deepen the concentration on a single supplier of a product South Africa does not make.[4]