The protein system
Thirteen states, about 171 million people, and a protein base built from starch. Maize output was 13.74 million tonnes in 2023 and rice 5.5 million, with cassava at 32.7 million tonnes in 2024; cereals, roots and tubers supply 81 per cent of dietary energy in Madagascar and 68 in Zambia.[2][3] Animal protein comes from extensive systems: 32.6 million cattle, 39.2 million goats and 218 million chickens yield 1.33 million tonnes of beef and 2.06 million of cow milk, with capture fisheries carrying much of the rest in Angola, Zambia, Mozambique and Comoros.[3]
The system is therefore cheap in resource terms and thin in nutritional terms. Only 555,000 of 4,475,000 tonnes of domestic protein supply went to animal feed in 2023, 12.4 per cent, because little protein passes through a feed conversion step.[1] Animal protein averages 13.9 grams a day, 27 per cent of the total, on an internal spread wider than that between most world regions: Madagascar 5.0 grams against Seychelles 56.4.[1] Trade is small in aggregate and decisive at the margin: gross protein import dependency was 19.6 per cent, 9.1 net,[1] and South Africa supplied 36.6 per cent of the tonnage and 36.2 of the value of a USD 10.82 billion agri-food import bill.[4]
Durable domestic capacity: 48
Quantitative score 53, adjusted down 5 points
The headline ratio flatters the region. Our computation from the FAO Food Balance Sheets puts protein self-sufficiency at 91.9 per cent in 2023, up from 89.3 in 2022, better than most entities in the index.[1] The indicators put the pillar at 53; we deduct five because the ratio has a suppressed denominator, taking it to 48: FAO scores regional dietary energy supply at about 103 per cent of requirement, Madagascar 89 and Zambia 93.[2]
Instability is the graver finding. Regional maize fell 17.9 per cent to 11.28 million tonnes in 2024, with Zambia down 54 per cent, Namibia 59, Lesotho 56, Botswana 45 and Zimbabwe 38.[3] Undernourishment averages 26.7 per cent across the eleven members with data, from 4.9 in Mauritius to 41.8 in Madagascar.[2] Deliberate investment has worked where it happened, Zambian aquaculture rising from 10,530 tonnes in 2011 to 76,627 in 2023 and Zimbabwe's irrigated wheat reaching 639,942 tonnes in 2025.[24][29] Both remain exceptions, and inland fisheries are being drawn down, with Lake Malawi's chambo populations sharply reduced.[27]
Resource headroom: 55
Quantitative score 66, adjusted down 11 points
This is the region's strongest pillar: 29.6 million hectares of arable land, 216.4 million of pasture, 1.27 hectares of grazing per person.[6] Water stress under SDG 6.4.2 runs at 0.3 per cent in Namibia, 1.8 in Angola and Mozambique and 8.6 in Zambia, against 67.6 in South Africa.[7] Nutrient application averages 25.5 kilograms of NPK per hectare against a world average of 117, and 2.0 in Comoros, so intensification headroom alone is large.[5][8]
The stock is real; the trend and the record of mobilising it are weaker. Arable land grew 31.3 per cent between 2000 and 2024, but regional population grew 89.9 per cent over the same period, so arable land per person fell 30.9 per cent, from 0.244 to 0.169 hectares, and the irrigated share of arable land fell from 6.47 to 5.76 per cent because irrigation investment did not keep pace with even the rainfed expansion that occurred.[6] Only 1.71 million hectares are equipped for irrigation today, 5.8 per cent of arable land, so low water-stress readings describe unbanked water that arrives as rain or does not arrive at all.[6] Agricultural credit tells the same story unevenly: Zambia's orientation index of 3.64 shows a functioning market, Namibia 2.49 and Zimbabwe 1.72 are moderate, and Mozambique, the region's largest arable-land holder at 8.28 million hectares, scores 0.064, a figure that has nearly halved since 2019; Eswatini, Madagascar, Mauritius and Comoros carry no series at all.[58] Two flagship expansion programmes make the same point in the field: Mozambique's ProSAVANA targeted 11 to 14 million hectares of the Nacala corridor before its 2020 termination, and Zambia's Farm Block Development Programme has run for two decades without finding an anchor investor for its 110,000 hectare Nansanga block, disbursing only 14.7 per cent of the funds allocated to it between 2022 and 2024.[59][60][61] Fisheries are where headroom has already gone: Namibia's hake resource is judged significantly below the biomass supporting maximum sustainable economic yield, the 2026 quota set at 133,000 tonnes and horse mackerel cut 5 per cent to 197,000.[25][48] Indian Ocean tuna, which underpins the islands, is healthier, with only bigeye assessed as overfished.[26]
Import exposure: 50
Quantitative score 56, adjusted down 6 points
Gross exposure is modest and concentration severe. Zambia is a net cereal exporter and Malawi imports 1.9 per cent of its cereals, yet six members buy essentially all their maize from one country: 100 per cent for Lesotho, 99.9 Botswana, 99.8 Namibia and Eswatini, 98.3 Mozambique, 97.4 Zimbabwe, a maize supplier Herfindahl of 7,347.[2][4] On food import value the indices reach 8,629 for Botswana and 8,523 for Lesotho, and dairy is 73.2 per cent South African, while wheat is diversified at 997.[4]
The indicators put the pillar at 56; we deduct six, taking it to 50, because a Herfindahl index treats South Africa as a supplier substitutable at a price, and it is not: it sits inside the same rainfall system as its customers, so the dominant source fails in the season regional harvests fail. Its own 2023/24 crop fell 22 per cent to 12.85 million tonnes on the same drought, while Zimbabwe absorbed 57 per cent of South African maize exports, some 1.3 of 2.2 million tonnes to April 2025.[16] The regional preference for white maize narrows the serious supplier set to South Africa and Tanzania.[14][17] The 2025-26 foot-and-mouth outbreaks constricted the same channel for animal protein, with outbreaks confirmed in five states including South Africa, which lost its FMD-free status and saw export suspensions cut its trade while the region kept depending on its grain.[32]
Maritime chokepoint exposure is low; landlocked members depend instead on the Durban, Beira, Nacala, Walvis Bay and Dar es Salaam corridors,[39] while the islands import 78 to 100 per cent of their cereals.[1][2]
Upstream dependence: 54
Quantitative score 48, adjusted up 6 points
Dependence is nominally severe and materially moderate. The region used 803,029 tonnes of fertiliser nutrients in 2023 and produced 48,900, 6.1 per cent, importing 984,285 tonnes, with domestic nitrogen confined to Zimbabwe's 5,500 tonnes.[5] Every member except Angola is a net fuel importer, poultry and dairy genetics are imported, and in Malawi and Zambia input subsidy programmes are the largest public agricultural investments, at several per cent of GDP a year.[38]
The indicators put the pillar at 48; we add six, taking it to 54, because those shares overstate the exposure that matters. At 25 kilograms of NPK per hectare against a world average of 117, and under 10 in Angola, Mozambique, Madagascar, Lesotho and Comoros, an input price shock compresses yields rather than halting production.[5][8] The 2022-23 shock is the test: Angolan nitrogen use fell 41 per cent and phosphate 68 without output collapsing.[5] Feed protein is partly regional, 1.17 million tonnes of soybeans grown in 2023 and Zambia expecting about 300,000 tonnes of soymeal from its 2025/26 crop at a moderate supplier concentration of 1,603,[3][4][36] though thin competition in Zambian and Malawian crushing keeps feed costs high.[37] Foot-and-mouth vaccine is made inside the region, below SADC requirement.[34] Labour is domestic and abundant, 73.0 per cent of Mozambican employment.[11] Energy is the sharpest residual risk because it is coupled to the same rainfall: Kariba usable storage fell to 13 per cent in 2024 with prolonged load-shedding in Zambia and Zimbabwe, and Zimbabwe's ZETDC ring-fenced 150 MW to irrigate winter wheat.[30][31][49]
Access and affordability: 18
Quantitative score 18, no adjustment applied
This is the region's weakest pillar and the binding constraint on the overall score. The affordability indicator, the population-weighted share of people unable to afford a healthy diet, is 82.3 per cent, from 15.2 in Mauritius to 93.6 in Madagascar, against a world figure of 31.9 per cent,.[52] The secondary indicators confirm the bottom of that band rather than the top. Undernourishment is 26.7 per cent population-weighted, more than three times the global average of 8.2 per cent.[2] Child stunting is 36.2 per cent population-weighted, above the WHO very-high threshold of 30 per cent, and reaches 47.7 in Angola, albeit on a wide confidence interval reflecting the absence of a recent survey there.[53] Food price volatility is severe on average, population-weighted food CPI rising 18.5 per cent in 2025 against 2024, though Zimbabwe's headline 42.2 per cent is a currency story that had collapsed to 0.7 per cent year-on-year by the first quarter of 2026, while Malawi's inflation held near 21 per cent into the same quarter.[54] No FAO GIEWS price warning was active for any of the thirteen members in July 2026, so the pressure the pillar measures is chronic cost rather than acute market failure.[56] Physical access is degraded in specific pockets rather than uniformly: 474,410 people remain displaced in Mozambique's Cabo Delgado as at February 2026, and Madagascar's October 2025 military takeover sits alongside 1.57 million people in acute food insecurity nationwide, projected to reach 1.8 million, with Crisis and isolated Emergency conditions concentrated in the Grand Sud.[50][42][43][44]
We apply no adjustment. Nothing among the admissible grounds pushes convincingly in one direction: the case that a market-based unaffordability measure overstates exposure for subsistence smallholders is undercut by Malawi's own figures, where 102.5 per cent protein self-sufficiency[1] coexists with 91.5 per cent unable to afford a healthy diet and 58.1 per cent severely food insecure.[2][52]
Shock endurance: 39
Quantitative score 35, adjusted up 4 points
This is one of the region's two weakest pillars. Import cover is thin to non-existent: Zimbabwe 0.52 months, Eswatini and Malawi 1.91, Zambia 3.29, Mozambique 3.44.[9] GNI per capita is USD 560 in Madagascar, 570 in Mozambique and 600 in Malawi, leaving no capacity to outbid on world markets,[10] and severe food insecurity is 24.6 per cent population-weighted, 58.1 in Malawi.[2] The 2024 El Nino is the documented test and the system failed it: February 2024 was the driest in a century at about 20 per cent of normal rainfall, over 30 million people were affected, six states declared a state of emergency, and the appeal reached USD 5.5 billion.[13][46] Only South Africa and Tanzania left that season in surplus.[14]
The indicators put the pillar at 35; we add four, taking it to 39, for two capacities reserve metrics miss. First, household substitution breadth: with 52 to 73 per cent of the workforce in agriculture in the six largest members, and cassava, sorghum, millet, pulses and small livestock alongside maize, a price shock leaves most of the population closer to food than an urbanised importing economy would.[3][11] Second, demonstrated rebound: Zambian maize went from 1.51 million tonnes in 2024 to a forecast 4,937,605 for 2025/26, and Zimbabwe's Cabinet expects a reserve above 900,000 tonnes from the same season.[18][20][47] Neither offsets the absence of fiscal depth, and the external absorber has gone: the World Food Programme closed its Southern Africa bureau, now run from Nairobi, after termination of about 90 per cent of USAID contracts.[15]
Crisis purchasing power scores 45; the indicators put the pillar overall at 35. General government debt reaches 102.5 per cent of GDP in Mozambique and 86 per cent in Zambia even after its 2024 Eurobond restructuring; Zimbabwe's comparatively low 43.8 per cent ratio reflects exclusion from international capital markets by arrears rather than fiscal headroom, and Comoros's lowest-in-region 29.5 per cent reflects aid dependence rather than market strength.[57] Market access to buy through a shock is narrow almost everywhere in the region, and the one social-protection system that had operated at regional scale, the World Food Programme, has just been cut by nine-tenths, with no member's domestic safety net carrying a verified coverage figure to fill the gap.[15]
Trajectory
Reserves are being rebuilt from two good harvests: Zambia's Food Reserve Agency is buying at least 500,000 tonnes of maize for 2026/27, Zimbabwe holds 500,000 tonnes of strategic capacity against an ambition of 1.5 million, and Malawi's National Food Reserve Agency has 180,000 tonnes of silos.[19][21][22] Malawi's agency was nonetheless still awaiting funds in June 2026 for a planned 108,000 tonne purchase. Storage exists; financing does not.[23]
Against that, FEWS NET reports an El Nino ongoing and strengthening into the 2026/27 season, with Crisis (IPC Phase 3) projected for southern Angola, southern and central Malawi, southern Mozambique and the Grand Sud of Madagascar in the October 2026 to January 2027 lean season, and isolated Emergency pockets among Angolan agropastoralists without livestock.[12] NOAA had an El Nino Advisory in force in June 2026, the Nino 3.4 index at 0.98 degrees C,[51] reversing the favourable outlook SARCOF-31 issued for 2025/26.[45] Foot-and-mouth disease is simultaneously disrupting intra-regional livestock trade and, for Botswana, EU access to veterinary zone 3c.[32][33]
What would change this picture is narrow. Irrigation carries the most leverage, since 5.8 per cent equipped area, a share that has fallen rather than risen since 2000, is what turns a rainfall anomaly into a national disaster,[6] and raising nutrient rates towards half the world average of 117 kilograms per hectare would lift the protein floor without new land.[5][8] Credit is the mechanism behind both: Zambia's own Farm Block Development Programme and Mozambique's ProSAVANA show that designating land is not the same as financing it, and Zimbabwe's shift to bankable title deeds in 2025 is the clearest live test of whether tenure reform can unlock the credit that irrigation and intensification both require.[58][59][60][62][63] Diversifying white maize supply away from South Africa, and funding a regional reserve with capital rather than ministerial intent, would address the correlated-failure risk both our imports and endurance adjustments penalise. Affordability carries the most leverage of all: with 82.3 per cent of the region's people unable to afford a healthy diet, no change to production or trade volumes changes this picture much unless it also reaches household incomes and food prices. On present policy, the region's trajectory through 2026/27 turns on whether affordability gains reach households rather than on production or trade volumes.
Country notes
Angola. The largest population at 36.7 million and the most self-financing importer, oil revenue behind a USD 2.47 billion food bill and the region's most diversified supplier base, Herfindahl 664.[4] Protein self-sufficiency of 78.2 per cent is the lowest of the large mainland producers, on 11.9 million tonnes of cassava and near-nil fertiliser use.[1][3][5] FEWS NET gives it the worst forward outlook.[12]
Zambia. Both the volatility and the upside: 122 per cent protein self-sufficiency and net cereal exports beside 41.0 per cent undernourishment.[1][2] Maize fell to 1.51 million tonnes in 2024 and is forecast at 4,937,605 tonnes for 2025/26.[3][47] It has the region's best intensification, the region's strongest agricultural credit orientation index at 3.64, and the largest tilapia industry,[5][28][29][58] yet its own two-decade Farm Block Development Programme has left its flagship Nansanga block without an anchor investor and disbursed only 14.7 per cent of the funds allocated to it between 2022 and 2024,[60][61] against 3.29 months of cover and Kariba's rainfall coupling.[9][30]
Malawi. The most agronomically self-contained member and among the most food insecure: 1.9 per cent cereal import dependency and 102.5 per cent protein self-sufficiency, yet 58.1 per cent severely food insecure.[1][2] The constraints are money and land: USD 600 GNI per capita, 1.91 months of cover, 0.19 hectares of arable land per person.[6][9][10]
Mozambique. Large, coastal and among the least nourished, at 44.7 grams of protein and 8.7 animal, with 47.3 per cent cereal import dependency and 98.3 per cent of maize from South Africa.[1][2][4] Its assets are 8.28 million hectares of arable land, the region's largest, and the Beira and Nacala corridors;[6][39] that land carries the region's weakest agricultural credit orientation index, 0.064 and falling, and its own flagship expansion programme, ProSAVANA, targeted 11 to 14 million hectares of the same corridor before its 2020 termination.[58][59] Against them sit USD 570 GNI per capita and 474,410 displaced in Cabo Delgado as at February 2026.[10][41][50] The LNG restart feeds nobody this season.[40]
Zimbabwe. The most policy-active and least financially resilient. Animal protein of 31.3 grams is the mainland's highest and irrigated wheat reached 639,942 tonnes in 2025,[1][24] yet reserves cover 0.52 months and it took 57 per cent of South African maize exports to April 2025.[9][16] Land redistributed since 2000, some 11 million hectares taken from about 4,500 commercial farmers, cut output of the four main large-scale commercial crops 30 to 70 per cent, and beneficiaries held it under non-bankable 99-year leases for two decades; a 2025 reform is issuing title deeds to 383,500 of them, though Parliament has questioned whether the new document is genuinely bankable while the state retains ownership.[62][63] A 2026 foot-and-mouth outbreak in Matabeleland South threatens the livestock base.[32]
Botswana. Structurally the most import-dependent mainland state: 36.4 per cent protein self-sufficiency, 82.7 cereal import dependency, food import Herfindahl 8,629.[1][2][4] Only 260,000 hectares are arable, so beef is the base, and the herd fell to 918,057 head in 2024.[3][6] Outbreaks from January 2026 cost it EU access for veterinary zone 3c under Implementing Regulation (EU) 2026/451.[33]
Namibia. A protein exporter and near-total grain importer: 51,200 tonnes of fisheries protein against 3,800 tonnes of domestic fish supply, and 22.70 million kilograms of beef exported in 2024, 13.25 million to the EU.[1][35] Self-sufficiency of 92.1 per cent sits beside 69.4 per cent gross protein import dependency, a trading structure, and maize fell 59 per cent in 2024.[1][3]
Lesotho. Functionally an extension of the South African food system: 92.3 per cent of food import value and 100 per cent of maize come from South Africa.[4] Protein self-sufficiency is 45.5 per cent with the region's highest dairy dependence, the cattle herd fell 23 per cent and maize 56 in 2024, and food imports take 43 per cent of export earnings.[1][2][3]
Eswatini. The most water-constrained member at 77.56 per cent water stress, above South Africa's 67.6, though 28.2 per cent of arable land is equipped for irrigation, the region's highest share.[6][7] Protein self-sufficiency fell to 35.1 per cent in 2023, the lowest of the thirteen, on 1.91 months of cover.[1][9] It has foot-and-mouth outbreaks of its own while relying on South Africa for grain.[32]
Madagascar. The worst nutritional outcome and the sharpest divergence from the regional score: 39.5 grams of protein, 5.0 animal, energy adequacy 89 per cent, undernourishment 41.8, alongside 92.9 per cent protein self-sufficiency on 4.97 million tonnes of rice. That is self-sufficiency at a subsistence floor.[1][2][3] A military takeover in October 2025 removed the government, and UN reporting in February 2026 put 1.57 million people in acute food insecurity, projected to reach 1.8 million, including 84,000 in emergency conditions.[42][43][44]
Mauritius. The most import-dependent state at 90.9 per cent cereal import dependency and 32.9 protein self-sufficiency, alongside the region's best nutrition, 105.0 grams of protein and undernourishment of 4.9 per cent, which measures purchasing power.[1][2] Tuna processing for the EU is the protein industry,[26] and 5.16 months of cover on USD 14,040 GNI per capita buys it through a shock.[9][10]
Seychelles. A statistical outlier: 15,400 tonnes of fisheries protein against 900 tonnes of domestic fish supply, so the balance sheet shows self-sufficiency above 240 per cent and import dependency above 190 at once, as tuna is landed, canned and re-exported.[1] It grows no cereals at all, and sovereignty rests on IOTC management and 3.67 months of cover.[9][26] It is excluded from the regional aggregate.
Comoros. The most financially exposed member: food imports equal 243 per cent of merchandise export earnings, so remittances and aid pay for food.[2] Protein self-sufficiency is 53.8 per cent with 27.4 per cent severely food insecure, yet protein supply is 63.0 grams from artisanal fisheries and a dense pulse-growing smallholder base.[1][6] Its 7.61 months of cover, the region's highest, rests on external transfers.[9]