The protein system
Vietnam feeds 101.6 million people[11] a diet that is broad and protein-rich for its income level. Protein supply reached 102.2 grams per person per day in 2023, above Japan's 93.8 grams in 2022 and well above Indonesia at 77.8, the Philippines at 77.5 and Thailand at 68.6 in the same year.[5] Of that, 43.2 grams comes from animal sources, 42 per cent of the total.[4] The largest single contributor is domestically grown rice at 27.0 grams, followed by pigmeat at 13.3 grams, fish and seafood at 11.1 grams, equivalent to a seafood supply of 40.1 kilograms per person per year,[9] and poultry at 8.8 grams.[4] Undernourishment is 5.3 per cent and moderate or severe food insecurity 10.7 per cent,[25][26] both low for the region.
Production matches consumption closely in most categories. In 2025 Vietnam produced 3.94 million tonnes of pork, 1.43 million tonnes of chicken meat and 285,000 tonnes of beef, covering 97, 90 and 68 per cent of domestic consumption respectively.[3] Paddy output was 43.5 million tonnes and rice exports 7.9 million tonnes.[12][1] Fisheries and aquaculture output reached 9.95 million tonnes on national accounting, of which aquaculture supplied 6.1 million tonnes;[14] FAO's comparable series for 2023 records 3.44 million tonnes of capture and 5.40 million tonnes of aquaculture.[6][7] Dairy is the single conspicuous gap, with 2.96 million tonnes of milk imported against 1.25 million tonnes produced in 2023.[4]
The counterflow is feed. Vietnam imported 13 million tonnes of maize, 5.7 million tonnes of soybean meal, 3.2 million tonnes of soybeans, 5.8 million tonnes of wheat and 1.5 million tonnes of distillers' dried grains in the 2025/26 marketing year,[1][2] with FAO recording total cereal imports of 22.4 million tonnes, well above the long-run average.[12] Domestic output of 22 million tonnes of animal feed and 6.5 million tonnes of aquafeed sits on top of that stream.[1]
Durable domestic capacity: 70
Quantitative score 74, adjusted down 4 points
We calculate an animal-protein self-sufficiency ratio of 107 per cent for 2023 from FAO Food Balance Sheets, against 78 per cent for total dietary protein once imported feed grain is counted into domestic supply.[4] Vietnam sells more animal protein abroad than it buys, and the gap between the two ratios is the imported feed.
The durability question is about the resource base rather than current output. Capture fisheries have flattened at about 3.4 million tonnes and the state's own Fisheries Development Strategy targets a reduction to 2.8 million tonnes by 2030, down from 3.8 million tonnes in 2020.[6][24] Peer-reviewed modelling projects catch yields falling 35 to 46 per cent by mid-century.[27] The European Commission's IUU yellow card has stood since 2017 and remains unresolved, with enforcement and traceability outstanding.[23] Rice area is contracting from 7.02 million hectares in MY2024/25 to a forecast 6.71 million in MY2026/27, and maize area is in continuous retreat as farmers move to fruit and vegetables.[1]
Two things the ratios cannot show justify a downward adjustment of four points. African swine fever is now endemic: roughly 1.2 million hogs were culled in 2025,[1] and 6.59 million pigs died or were culled across 17,564 outbreaks between 2019 and 2024,[18] a recurring annual tax on the pork base. Second, the Mekong Delta, which anchors both rice and pangasius, is degrading from causes largely beyond Vietnam's control. Salinity reached 30 to 45 kilometres inland in the 2025 dry season and occasionally 60 to 70 kilometres,[20] and water at four grammes per litre, four times the tolerance of most crops, reached 15 to 20 kilometres up most delta rivers in 2024-25.[19]
Resource headroom: 50
Quantitative score 44, adjusted up 6 points
Vietnam has very little land headroom. Arable land is 0.067 hectares per person,[11] among the lowest of any large agricultural producer, and has fallen 19.1 per cent per person since 2000 even as the broader cropland base, which includes rubber, coffee, fruit and cashew, grew fast enough to add 7.3 per cent per person despite population growth of 30.9 per cent over the same period.[37][39] The expansion landed on permanent tree and shrub crops, not on the annual cropland that carries rice, maize and feed grain, which grew only a sixth as fast. Water looks more comfortable in aggregate, at 3,606 cubic metres of internal renewable resources per person per year with total withdrawals at only 22.8 per cent of that, though agriculture already accounts for 94.8 per cent of withdrawals.[11] Irrigation infrastructure is extensive rather than idle: 68.1 per cent of arable land was equipped for irrigation in 2024, up from 60.8 per cent in 2000, the highest reading of any entity we have tested against this measure, though the absolute equipped area has been reported unchanged at 4.585 million hectares in every year since 2008 and, measured against the faster-growing cropland base, irrigated coverage actually fell from 48.7 to 39.4 per cent.[37] That comfort is partly illusory in another sense too: the Mekong and Red River systems are both transboundary, and delta salinity is projected to expand by 10 to 27 per cent by 2050 from human pressures alone, with sea level rise adding 5 to 19 per cent.[21]
The protein mix is feed-hungry. Pigmeat and poultry supply 54 kilograms per person per year,[8] driving feed demand from 28.62 million tonnes in 2025 to a forecast 30.60 million in 2027[1] while domestic maize covers about a quarter of consumption and shrinks.[1] Wild fish are an input as well as an output: 1.86 million tonnes went to feed use in 2023,[4] drawn from a capture sector the state intends to contract.
Capital is not the constraint. Agricultural credit reached USD 40.8 billion in 2024, nearly tripled from USD 14.3 billion in 2012, though its share of a faster-growing total credit market fell from 9.6 to 6.6 per cent over the same period and the agriculture orientation index sits below parity at 0.575.[38] Institutional delivery is demonstrated rather than aspirational: the Mekong Delta's One Million Hectares low-emission rice programme cleared its 2024-25 interim target of 180,000 hectares at roughly double the planned pace, reaching more than 354,000 hectares by the end of 2025 with participating farmer income up 13.4 per cent and emissions cut by three to four tonnes of CO2 equivalent per hectare against conventional cultivation.[40]
We add six points for aquaculture. Pangasius and tilapia are freshwater pond systems with low marine-ingredient inclusion, aquaculture fish and shrimp output rose 5 and 6 per cent in 2025,[2] and installed aquafeed capacity of about 5.5 million tonnes runs at only 60 to 70 per cent utilisation.[22] This is a funded, operating capability converting feed into protein at materially lower land and water cost than terrestrial livestock, and land and freshwater statistics do not see it. The credit and delivery evidence above corroborates that this kind of claim is mobilisable for Vietnam specifically: agricultural credit is real, large and growing in absolute terms, and the country's flagship land programme shows that mobilisation capacity converts into delivered, measured outcomes.
Import exposure: 66
Quantitative score 70, adjusted down 4 points
Direct protein import exposure is low, because Vietnam's dominant protein sources are domestic. Net animal-protein import dependency was minus 5.3 per cent of domestic supply in 2023.[4] Meat and edible offal imports of 978,300 tonnes worth USD 2.004 billion in 2025[17] equal roughly 15 per cent of the 5.65 million tonnes of pork, chicken and beef produced at home,[3] and their supplier base is wide, with India the largest origin at 19.3 per cent of volume.[17] Wheat sourcing is diversified, with Australia at 26 per cent, Brazil 19, the United States 17 and Ukraine 13 in 2025.[1] Route exposure is real without being binding: South American cargoes pass the Malacca Strait, which has usable substitutes in Sunda and Lombok, and North American cargoes cross the Pacific unobstructed.
We deduct four points for two exposures outside the headline ratios. Dairy is about 69 per cent imported by volume with no credible domestic substitution path at scale,[4] and the meat import bill grew 12.2 per cent in a single year while endemic African swine fever caps domestic pork growth,[17] so the import share is rising rather than stable. Vietnam is also an exceptionally open economy, with imports of goods and services equal to 92.1 per cent of GDP in 2025,[11] so a general trade shock reaches the food system through many channels at once.
Upstream dependence: 37
Quantitative score 32, adjusted up 5 points
Upstream dependence is what keeps Vietnam's tier down. Domestic feed supply is almost entirely by-products: 2.8 million tonnes of rice bran and broken rice, 1.7 million tonnes of maize and 550,000 tonnes of cassava, totalling 5.05 million tonnes against 23.57 million imported.[1] Vietnam grew 43,000 tonnes of soybeans in 2025 against imports of 3.2 million tonnes of beans and 5.7 million tonnes of meal.[2] The maize and soybeans imported in 2023 carried about 1.39 million tonnes of protein, equal to 79 per cent of the 1.77 million tonnes of animal protein produced that year.[4]
Concentration compounds the volume. Argentina shipped 4.18 million tonnes of soybean meal to Vietnam in 2024, worth USD 1.60 billion and roughly 69 per cent of the total.[15] Argentina and Brazil together supplied 82 per cent of maize imports in 2025;[1] partner-reported data put Argentine maize shipments at 6.81 million tonnes and Brazilian at 4.64 million in 2024.[16] The United States supplied 72 per cent of DDGS.[1] Fertiliser is the counterweight: Vietnam produced 1,798,315 tonnes of nutrient nitrogen in 2023 against agricultural use of 1,764,701 tonnes, making it a net exporter, and covered 84 per cent of phosphate use.[10] Potash is the exception, with no domestic K2O production recorded in 2023 against 491,555 tonnes of nutrient imports.[10]
We add five points. The fertiliser position means the crop side of the input chain is largely sovereign, and two new soybean crushing lines commissioned in the first quarter of 2026 lift domestic meal output from 1.48 million tonnes in 2025 towards a forecast 2.34 million tonnes in 2027,[2] shifting the dependency from finished meal to raw beans and widening the supplier field. Both are operating rather than announced.
Access and affordability: 76
Quantitative score 79, adjusted down 3 points
Access is not the binding constraint. The FAO/World Bank Cost and Affordability of a Healthy Diet series puts the share of Vietnam's population unable to afford a healthy diet at 7.9 per cent in 2024, down from 11.1 per cent in 2017.[28] The path was uneven, rising back to 9.6 per cent in 2021 as the pandemic hit incomes before resuming its decline.[28] We score toward the foot of the band because the outcome checks below are weak. Undernourishment is low, at 5.3 per cent of the population in 2023,[25] consistent with adequate aggregate supply.
Child stunting is the weak outcome check. The current Joint Child Malnutrition Estimates series puts stunting at 18.2 per cent of children under five in 2023, down from 24.6 per cent in 2015,[29] and Vietnam's own 2023 national nutrition survey reported 18 per cent.[30][31] Both sit well above high-income peers, and the pace of reduction has slowed to under one percentage point a year since 2015.[30] Price volatility is episodic rather than chronic: the food and catering services CPI group rose 4.03 per cent in 2024, contributing 1.35 percentage points to a 3.63 per cent headline rate,[35] with rice prices up 15.93 per cent in the same year on export-driven swings,[32] and Vietnam's food inflation series has historically ranged from plus 74.29 per cent in June 2008 to minus 10.12 per cent in July 2009, though recent readings sit near 4.6 to 4.9 per cent.[33]
Physical access is the sharpest divide within the pillar. The northern midlands, mountains and Central Highlands, where ethnic minority populations are concentrated, carry stunting rates of about 30 per cent, well above the national figure,[31] pointing to weaker market reach and infrastructure rather than a national supply shortfall. The national averages therefore understate the affordability and nutrition problem facing Vietnam's poorest and most remote households. Undernourishment and stunting are Vietnam's weakest outcome indicators, the latter at 18.2 per cent of children under five in 2023,[29][36] which puts the indicators at 79. The regional gap sits inside the pillar's physical-access component, which the band floor prevents from lowering the score any further, so we deduct 3 points instead, on the distributional ground the rubric sets out, taking the pillar to 76.
Shock endurance: 56
Quantitative score 49, adjusted up 7 points
Physical cover is thin. At MY2025/26 ending stocks, rice covers about 43 days of consumption, wheat 39, maize 15 and soybean meal 14,[1][2] so a feed shipping interruption would reach livestock and aquaculture output within weeks.
Against that stand two real buffers. Rice exports of 7.9 million tonnes equal 35 per cent of domestic consumption and can be retained,[1] a very large and fast-acting reserve, and the USD 11.34 billion seafood flow could be redirected inward.[13] Dietary breadth is good, with rice, pork, poultry, freshwater fish, marine fish, crustaceans and eggs all material, and 26 per cent of protein from domestic rice.[4]
Crisis purchasing power scores 35, the weakest of the pillar's components; the indicators put the pillar overall at 49. FX reserves of USD 85.58 billion against an import bill equal to 92 per cent of a USD 514 billion economy give only about 2.2 months of cover,[11] below the conventional three-month adequacy threshold even though the absolute reserve stock is sizeable, and GDP per capita of USD 5,066 leaves households price-sensitive.[11] Market access sits just below investment grade, with Moody's at Ba2 on a positive outlook and S&P at BB+ on a stable outlook,[34] so Vietnam could borrow through a shock but at a rising premium; no verified figure for social-protection scale-up capacity could be established.
We add seven points for documented shock performance. African swine fever destroyed roughly six million pigs, a quarter of the national herd, in 2019, and pork output fell from 3.43 million tonnes in 2018 to 2.93 million in 2020; production regained the pre-outbreak level by 2023 and reached 3.94 million tonnes in 2025, 15 per cent above the pre-ASF peak, while the sector restructured towards biosecure commercial units now supplying 62 per cent of output.[3][18] Substitution was visible again in 2025, with consumers switching from pork to chicken and feed mills substituting DDGS for maize.[1]
Trajectory
The direction of travel is more protein output and no less upstream dependence. Feed demand is forecast to rise 7 per cent between 2025 and 2027 while domestic maize output falls,[1] so the import bill grows structurally. Two levers count for more than everything else. The first is domestic crush and co-product capacity: the 2026 crushing expansion is the right shape of intervention, and extending it alongside cassava, rice bran and novel protein routes would cut a finished-meal dependency concentrated in one country. The second is aquaculture composition, since a tonne of pangasius or tilapia protein carries a far lower land, water and feed-protein burden than a tonne of pork, and 1.6 to 2.2 million tonnes of aquafeed capacity already sits idle.[22]
Two risks would push this the other way. Losing delta productive area faster than modelled, through salinity, subsidence and upstream flow regulation, would erode the capacity pillar at its foundation.[21] A supply interruption in Argentina, whether from drought, export tax policy or logistics, would test a system holding fourteen days of soybean meal cover with 69 per cent of supply from that one origin.[15][1] A continued IUU yellow card would not threaten supply, though it would cut the export earnings that pay for the imports.[23]