The protein system
Japan feeds 124 million people on 4.3 million hectares of farmland, down from 6.1 million at the 1961 peak, and on 0.0325 arable hectares per person, among the lowest ratios of any large economy.[12][15][21] Protein supply sits at 77.1 grammes per person per day,[22] modest for a G7 economy and below the global average. The diet remains meaningfully marine, with seafood supply of 22.0 kilogrammes per person per year,[23] well down from its historical peak, against an animal-protein share of dietary energy of about 7.6 per cent.[24]
Of 2,248 kcal per person per day in FY2024, domestic production supplied 860 kcal, giving the calorie self-sufficiency ratio of 38 per cent now held for a fifth year; the production-value ratio rose three points to 64 per cent, and a new intake-calorie measure from the 2025 Basic Plan reads 46 per cent against a 1,850 kcal reference requirement.[1][2] Rice is the single sovereign staple at 99 per cent, with wheat 16, soybeans 24 and seafood 47 on the same basis.[1]
Underneath sits a feed-driven distortion that separates production from sovereignty. As a domestic production ratio, which ignores where the feed came from, Japanese livestock products reach 65 per cent; as a self-sufficiency ratio, which excludes output attributable to imported feed, 17 per cent. The 48-point gap decomposes into beef at 50 against 14 per cent, pork 48 against 6, chicken 64 against 8, eggs 98 against 12 and dairy 64 against 29.[1][3]
The import side is large and, at national-calorie level, respectably spread. MAFF attributes 38 per cent of supply calories to domestic production, 23 to the United States, 12 to Australia and 8 to Canada, with Brazil, Malaysia, China and the European Union each at 2 to 4 per cent behind them. In value terms domestic production is 64 per cent and no foreign supplier exceeds 8.[1]
Durable domestic capacity: 35
Quantitative score 37, adjusted down 2 points
The indicators put capacity at 37; we deduct 2 points, taking the pillar to 35. The case rests on the self-sufficiency series above and the fisheries collapse beside it. Japanese fisheries and aquaculture produced 3,634,800 tonnes in 2024, marine capture accounting for 2,787,100 tonnes, a fall of 5.1 per cent and the lowest since the series began in 1951.[9] That is roughly 28 per cent of the 1984 peak of 12.8 million tonnes,[10] and edible-seafood self-sufficiency has fallen to a record 52 per cent.[11] Of the 50 stocks across 36 species assessed on a three-level scale in FY2023, 62 per cent were judged low, against 20 per cent high and 18 medium.[8]
We applied minus 2 because self-sufficiency ratios are static snapshots and Japan's producing base is not. Self-employed farmworkers fell 25.1 per cent in five years to 1.02 million in 2025, the largest decline recorded;[14] core agricultural workers numbered 1.1 million in 2024 at an average age of 69.2, with 71.7 per cent aged 65 or over;[13] paddy area fell from 2.11 million hectares in 1995 to 1.52 million in 2025.[12] The current 38 per cent carries a demographic clock.
Resource headroom: 40
Quantitative score 38, adjusted up 2 points
The indicators put headroom at 38; we add 2 points, taking the pillar to 40. Japan's headroom divides cleanly. Water is abundant, at 3,437 cubic metres of renewable internal freshwater per person per year,[16] which removes the constraint binding most low-arable protein importers and separates Japan decisively from the Gulf. Land is not, and the capture fishery has negative headroom: 62 per cent of assessed stocks are already low,[8] and marine aquaculture leans on a fishmeal stream that is roughly half imported, at 177,000 tonnes produced domestically against 179,000 imported.[6]
The protein mix is comparatively efficient by rich-country standards, weighted towards seafood, pork and poultry rather than beef, and beef consumption is falling further as prices outrun incomes.[7] That efficiency is undercut by where the feed grows: Japan will import an estimated 15.8 million tonnes of corn in MY2025/26, 12.5 million tonnes of it for feed, alongside 3.3 million tonnes of soybeans, 1.65 million tonnes of soybean meal and 1.13 million tonnes of barley.[5][6] Fertiliser is applied at 202.8 kilogrammes per arable hectare on land already worked hard.[20]
Two collected series test whether Japan can act on a retired-land reserve, and both confirm it can. Irrigation-equipped land covers 70.7 per cent of Japan's arable base in 2024, up from 58.6 per cent in 2020, so water infrastructure is not a constraint on land brought back into cultivation.[41] Japan also carries a full, unbroken agricultural credit series: lending to agriculture, forestry and fishing reached JPY 1.06 trillion in 2024, roughly five times its 2008 level, backed by the world's largest agricultural cooperative banking group and by the Japan Finance Corporation's dedicated agriculture unit.[42]
A third series narrows the size of the reserve itself. Cropland and arable land have both fallen every year since 2000, by 11.6 and 10.1 per cent respectively, with no net expansion anywhere in the record.[41] Of the 257,000 hectares officially classed as derelict farmland in 2024, only 94,000 hectares are judged restorable, well short of the 590,000-hectare paddy-exit figure, most of which moved into other actively farmed crops under the block-rotation subsidy programme rather than sitting idle.[43] The Farmland Bank, created in 2014 to consolidate scattered plots under larger operators, targeted 80 per cent of farmland held by core farmers by FY2023 under the 2013 Japan Revitalization Strategy; the actual figure was 58 per cent in 2021, 59.5 per cent in FY2022, 60.4 per cent in FY2023 and 62.1 per cent in FY2025, more than two years past the target date, though FY2023 alone added 21,000 hectares of newly consolidated farmland, up from 16,400 the year before.[43][44][45]
We applied plus 2 for a reserve that is real but smaller and slower than the raw farmland-exit figures suggest. Water and finance are not what is holding it back; a decade-old consolidation target still short by 18 points is.
Import exposure: 40
Quantitative score 34, adjusted up 6 points
The indicators put import exposure at 34; we add 6 points, taking the pillar to 40. Japan imports 62 per cent of its supply calories[1] and the great majority of the feed behind its animal protein, so the pillar has to score low. Concentration at product level is considerably worse than the calorie aggregate suggests. Corn, the largest single import by volume at 15.47 million tonnes in MY2024/25, came about 86 per cent from the United States.[5] Sorghum was 69 per cent American and 30 per cent Australian;[5] rapeseed 63.8 per cent Australian and 36.2 per cent Canadian; soybeans 68.3 per cent American and 21.1 per cent Brazilian.[6] Beef rests on two suppliers, Australia at 48 per cent and the United States at 35.[7] Pork is the best-hedged animal protein, at 28 per cent from the United States, 23 from Canada and 15 from Spain.[7]
We applied plus 6 on two grounds dependency ratios omit. First, chokepoint exposure is close to nil for the dominant flows: the United States, Australia, Canada and New Zealand supply 43 of Japan's 62 imported calorie points[1] and reach Japanese ports across the open Pacific, transiting none of Hormuz, Malacca, Suez, Bab el-Mandeb or the Bosphorus. Second, the five main classes of food wheat are state-traded, with MAFF as importer of record selling on to mills at a semi-annually set price, so 5.57 million tonnes of the most politically sensitive food grain moves under direct government control.[5] Offsetting both, these same sea lanes run through the Western Pacific, the theatre of the decade's most plausible great-power contingency, which caps the uplift at 6.
Upstream dependence: 30
Quantitative score 27, adjusted up 3 points
The indicators put upstream dependence at 27; we add 3 points, taking the pillar to 30. This is Japan's weakest pillar, and on this measure Japan is weaker than the Gulf Cooperation Council, which at least exports fertiliser and owns its energy. Close to 100 per cent of Japan's chemical fertiliser raw material is imported, and product-level concentration is severe: as of October 2024, 69 per cent of urea came from Malaysia and 99 per cent of ammonium phosphate from China.[25] Earlier analysis of MAFF data put nitrogen at 47 per cent Malaysian and 37 Chinese, phosphate at 90 per cent Chinese and 10 American, and potash at 59 per cent Canadian and 16 Russian.[26] A single strategically adversarial supplier therefore holds a near-monopoly over one of the three macronutrients. Feed is 74 per cent imported on a digestible nutrient basis,[1] with the concentrate share higher still, and energy powering the cold chain, protected horticulture and the fishing fleet was 87.3 per cent imported in 2023.[17]
We applied plus 3 for two owned capabilities. Japan imports no live cattle at all,[7] and its beef sector rests on domestically held Japanese Black genetics accounting for roughly half of cattle slaughtered and subject to export restriction, so one critical input class is fully sovereign. Domestic crushers also supply all of Japan's rapeseed meal demand and about half of its soybean meal demand,[6] so the processing step is owned where the raw material is not.
Access and affordability: 66
Quantitative score 68, adjusted down 2 points
The indicators put access and affordability at 68; we deduct 2 points, taking the pillar to 66. The headline figure has moved worse than a G7 economy at Japan's income level would be expected to show: 13.3 per cent of the population could not afford a healthy diet in 2024, a materially higher unaffordability rate than most peer G7 economies post.[32] Outcome checks are mixed: the World Bank series carries no undernourishment value for Japan at all, and we score the component at the sub-2.5 per cent floor by inference from income and outcomes rather than from a published number;[33] child stunting was 5.2 per cent in 2024, down from a peak of 6.9 per cent in 2008-09, though above South Korea (1.8 per cent), Germany (2.2), Australia (3.1) and the United States (4.2).[34]
Price stability, a historic Japanese strength, was tested rather than broken. Food price inflation has averaged 2.83 per cent since 1971,[35] but wholesale egg prices rose 72 per cent, from JPY 151 to JPY 260 per kilogramme, during the 2022-23 avian-influenza year,[28] and the FY2024-25 rice shortage forced a 590,000-tonne reserve release and a 14,635 per cent surge in private rice imports.[5] Physical distribution is excellent by any national measure, with a Logistics Performance Index of 3.9 of 5 in 2022,[36] though that aggregate measures the trade network rather than the last mile into a depopulating, ageing countryside, where MAFF's own commentary describes a food-access problem with no quantified national figure from a primary source.
We deduct 2 points because a national logistics score cannot see the rural access gap an ageing, depopulating countryside creates, a gap the affordability figure above does not capture. The deduction is limited to 2 points because the rural access-difficulty population has no quantified figure from a primary source, and because the affordability figure itself already carries the income-peer weakness directly.
Shock endurance: 75
Quantitative score 69, adjusted up 6 points
The indicators put endurance at 69; we add 6 points, taking the pillar to 75. This is Japan's second strongest pillar behind access and affordability, and it rests on money, law and latent land rather than deep reserves. Stocks are measured in weeks to months. The government rice reserve targets one million tonnes, sized to cover two consecutive poor harvests and replenished at about 200,000 tonnes a year.[5] Mills hold roughly 900,000 tonnes of imported food wheat, equal to 2.3 months of consumption, of which MAFF subsidises 1.8 months, and a further one million tonnes of imported feed grain sits under storage subsidy.[5]
Recent shocks were absorbed without rationing and with real stress. The 2022-23 avian influenza season produced 84 outbreaks and the culling of 17.7 million birds, 16.4 million of them layers, costing JPY 14.2 billion in direct farm losses and JPY 16.9 billion in state compensation,[27] with wholesale egg prices rising from JPY 151 to JPY 260 per kilogramme.[28] The FY2024-25 rice shortage was more instructive, because it hit a crop at 99 per cent self-sufficiency: MAFF released 590,000 tonnes of reserve rice and private rice imports rose 14,635 per cent to 88,706 tonnes.[5] Both episodes were domestic, so endurance against deliberate trade interdiction remains untested.
Crisis purchasing power is where Japan is least ambiguous. Foreign reserves stood at about USD 1.37 trillion in 2025,[19] equal to 14.0 months of import cover,[37] against a food import bill equal to 9.5 per cent of merchandise imports,[18] so Japan can outbid essentially any competitor for physically available supply. In 2020 the state distributed a universal, unconditional cash payment of JPY 100,000 per resident, about 42 per cent of a full-time worker's median monthly earnings, with no means test to administer, though municipal delivery was slow: the payment was announced in April and most households received it between late June and early July, and a mid-June survey of 43 large municipalities found three that had paid under a tenth of their residents.[40] The instrument is therefore proven in reach and unproven in speed. Set against that is a fiscal position that must be weighed rather than ignored: general government debt reached 248.7 per cent of GDP in 2025, the highest of any advanced economy, and Japan's 2020 stimulus, headlined at over 50 per cent of GDP, amounted to about 16 per cent of GDP in actual direct spending, still the third-largest in the G7.[38][39] Market access has so far stayed intact because the debt is overwhelmingly yen-denominated and domestically held, but this is a strength resting on reserves and currency status rather than on a fiscal position Japan could keep drawing on indefinitely.
We applied plus 6 for the substitution capability MAFF itself quantifies. Its food self-sufficiency potential index calculates that, with maximum use of existing farmland on a potato-centred cropping pattern, the domestic land base could supply about 2,418 kcal per person per day, against about 1,755 kcal under rice and wheat-centred cropping.[4] The higher figure sits above the 1,850 kcal reference intake MAFF now uses as the denominator of its intake-calorie self-sufficiency ratio.[1] This is a government-computed physical capability on land Japan already holds, reinforced by the 2024 Act on Countermeasures for Difficult Food Supply Situations, which grants staged statutory powers over production and distribution,[29] and by the June 2024 revision of the Basic Act, which made food security a founding principle.[30] The adjustment is capped at 6 because MAFF reports the potato-centred index declining on labour shortage, so conversion would be slow.
Trajectory
Japan's policy direction is now correct and its instruments remain modest. The revised Basic Act and the crisis-response statute give the state legal machinery it lacked in 2022.[29][30] The FY2020 Basic Plan target of 45 per cent calorie self-sufficiency by FY2030[31] has not moved the number in five years, which is why the 2025 plan added the intake-calorie indicator.[1] Fertiliser policy is the most concrete change, with targets to double sewage sludge and compost use by 2030 and raise domestically sourced phosphate fertiliser use to 40 per cent.[25] Against a 99 per cent Chinese share in ammonium phosphate, that is the highest-return intervention available to Japan.
Three developments would do the most to change Japan's exposure. Substituting a quarter of imported concentrate feed with domestic forage, silage maize or feed rice would lift capacity and upstream together, since the feed ratio is the hinge on which both turn. Delivering the phosphate and nitrogen substitution targets with funding attached would lift upstream out of its current band. Stabilising the capture fishery, where 62 per cent of assessed stocks are low,[8] would arrest erosion of the one protein Japan has historically produced for itself. Absent those, demography does the opposite: a farm workforce falling by a quarter every five years[14] will lower the ceiling faster than policy raises it.