SG
WELCOME TO
Singapore Food Security AI Command & Control Centre
Driving Food Resilience
Singapore imports more than 90 per cent of its food from over 180 countries and regions. This command centre scores food supply chain disruption risk continuously against that network — for the officers who hold the national instruments, for the enterprises who actually execute the substitutions, and, through AFSIM, for the region whose food system Singapore sits inside.
SELECT YOUR VIEW
VIEW 01
Government Agencies
The full national picture. Four-pillar command, FS Risk and the Malaysia dependency, source and corridor boards, stockpile coverage, agentic response staging, and the enterprise distribution tiers. 39 views.
ENTER NATIONAL C3 →
VIEW 02
Industry
The importer's own picture. Where your supply chain risks sit, what has disrupted your items before, how those disruptions can be predicted, and which approved alternate supplier to switch to when a source fails — across every food category SFA regulates. 19 views.
ENTER INDUSTRY PORTAL →
VIEW 03
ASEAN · AFSIM
The regional picture. AFSIM — the ASEAN Food Security Index & Monitor — measures, benchmarks and warns; the collective C3 beneath it holds interdependency, exportable surplus, correlated exposure, simultaneous substitution risk, reserves and coordinated response. Built on a verified data snapshot with a published register. 33 views.
ENTER AFSIM + ASEAN C3 →
SG
Singapore Food Security AIC³ + AFSIM v7.1
DRIVING FOOD RESILIENCE · ASEAN FOOD SECURITY INDEX & MONITOR
⇄ SWITCH VIEW
TOWER ACTIVE
SOURCE COUNTRIES
180+
OPEN SIGNALS
23
FS RISK
HIGH
AGENTS LIVE
15
COVERAGE DAYS
SGT
Command
FS Risk
Intelligence
Operations
Enterprise
Platform
Value
My Business
Food Categories
My Sources
Disruption
Act
AFSIM Command
Monitor
Network
Early Warning
Scenario & Response
Reporting
Data & Institutions
COMMAND › NATIONAL EXECUTIVE C3

National Food Security Command & Control

Independent prototype. This is a concept demonstrator built on granted Singapore patent SG 10202109991Y. It is not an SFA system and carries no SFA endorsement. Figures marked SFA are drawn from published Singapore Food Agency material; all others are marked ILLUSTRATIVE and exist only to demonstrate the operating model.

Singapore imports more than 90 per cent of its food, sourced in 2025 from over 180 countries and regions, up from about 140 two decades ago. That breadth is the country's principal defence — and it is only a defence if someone is watching all of it at once, continuously, and can tell which sources matter for which food group on any given day.

IMPORT DEPENDENCE SFA
>90%
Of food consumed is imported
SOURCE COUNTRIES SFA 2025
180+
Countries and regions supplying food
LOCAL PROTEIN SFA 2025
25%
Of protein consumed; 30% target by 2035
LOCAL FIBRE SFA 2025
8%
Of fibre consumed; 20% target by 2035
SOURCES IN ALERT ILLUSTRATIVE
7
Amber or red across food groups

Food Group Supply Assurance, Rolling 9 Months

Modelled probability that normal supply volumes for each food group can be maintained on the target date given current sanitary, climatic, regulatory and route conditions. Prototype series.

National Risk Posture by Domain

Command Priorities

Accelerate accreditation, poultry and egg
HPAI activity across multiple European and American sources; widen the approved set before the northern winter peak
P1
Pre-position rice agreement activation
Mekong climate signal elevated; agreements with Cambodia, Thailand and Vietnam are the instrument
P1
Fertiliser and freight watch
Strait of Hormuz disruption feeds through to yields with a two to three season lag
P2
Stockpile rotation review
Coverage across carbohydrates, protein and fibre against the current threat set
P2
Leafy vegetable surge readiness
Local farms are the buffer for the most perishable, least substitutable category
P2
Importer business continuity plans
Verify warm links to alternate sources are live, not nominal
P3

Five Command Functions

FUNCTION 01

Supply Radar

Monitors source countries, accredited establishments, routes and local farms for every food group.

Views 05 · 12 · 16 · 17
FUNCTION 02

Risk Engine

Applies the patented risk determination methodology. The single authoritative source of a risk value.

Views 03 · 21
FUNCTION 03

Demand Intelligence

National consumption by food group against confirmed supply, substitution headroom and stockpile cover.

Views 18 · 22
FUNCTION 04

Decision Engine

Converts a scored risk into which of the four pillars to pull, in what order, at what cost.

Views 03 · 04
FUNCTION 05

Agentic Operations

Prepares accreditation, industry engagement, stockpile and communication workflows for officer approval.

Views 26 · 28

Food Group State Board

Food groupImport dependenceConcentrationSubstitutabilityRecovery time if disruptedState
Rice & carbohydratesFully importedMekong region concentratedLow, variety-specificHarvests take months to years to recoverAMBER
Poultry & eggsEggs partly localHistorically Malaysia heavyMedium, form substitution worksFlock rebuild, monthsRED
PorkFully importedRegional live plus 20+ chilled and frozen sourcesHigh, chilled and frozenHerd rebuild, many monthsAMBER
Beef & muttonFully importedBroad, frozen-dominantHighMonthsGREEN
SeafoodPartly localRegional plus distant waterMediumWeeks to monthsGREEN
Leafy vegetablesLargely importedMalaysia and regional heavyLow, highly perishableWeeks, but quality degrades firstAMBER
Other vegetables & fruitLargely importedOver 100 sources for vegetables SFAHighWeeksGREEN
Processed & ambientLargely importedVery broadHighShelf life absorbs the shockGREEN
States are prototype assessments produced by the model, not SFA determinations. Structural characteristics — import dependence, concentration and recovery dynamics — follow SFA's published description of Singapore's food supply risks.
COMMAND › AI CONTROL TOWER

AI Control Tower

The brain and nerve centre. Two agent meshes operate around a single authoritative kernel: a sensing mesh that watches the world's food system, and an operating mesh that prepares national response options. Neither determines risk independently — the patented risk engine remains the sole authoritative risk-determination mechanism, and no agent may originate or override a risk score.

GLOBAL FOOD SUPPLY SCAN · 180+ SOURCE COUNTRIES
AI TOWER
AI Control Tower
SENSE · VERIFY · SCORE · REASON · ACT

Sensing Agent Mesh

Animal & Plant Health Watch
WOAH/WAHIS · IPPC · national veterinary services
Reads outbreak notifications and vernacular reporting from source countries; infers disease from clinical description before formal labelling, and tracks which zones remain eligible under regionalisation.
Trade Policy & Export Restriction
National gazettes · WTO notifications · ministerial statements
Detects export bans, quotas and price controls in source countries. The 2023 Indian non-basmati ban and the 2022 Malaysian chicken ban are the calibration cases.
Climate & Harvest
ECMWF · NOAA ENSO · crop condition indices
Watches Mekong hydrology, El Niño state, Malaysian and Cameron Highlands rainfall, and heat stress across producing regions.
Route & Maritime
AIS · chokepoint transit counts · war-risk premia
Monitors Suez, Bab al-Mandab, Hormuz and Malacca. Tracks air cargo capacity for perishables alongside sea freight.
Input & Commodity
Grain · feed · fertiliser · energy · freight rates
Upstream inputs move before food prices do. Feed and fertiliser are the two channels through which a distant conflict reaches a Singapore shelf.
Establishment & Consignment
Accreditation registry · import permits · inspection results
Watches the administrative layer: which establishments are accredited, which consignments are in transit, which failed inspection or testing.
Local Production
Farm returns · yields · productivity · farm count
Tracks the regenerative pillar: vegetable, egg and seafood output, farm-level productivity, and surge headroom in a crisis.
Adversarial Verification
Assert · falsify · source-reliability scoring
One agent asserts the event, a second attempts to falsify it against independent sources, a third scores source track record. In a national setting a false positive is a public confidence cost, not merely an operational one.

Operating Agent Mesh

Seven execution agents prepare the national response once the kernel has determined and classified the risk. In a government deployment these agents propose and stage; officers decide. Every agent is bounded by an authority ceiling and writes to the decision record.

Accreditation Agent
SOURCE WIDENING
Identifies candidate source countries and establishments for the affected item and assembles the accreditation dossier and priority order.
Industry Engagement Agent
IMPORTERS & ASSOCIATIONS
Identifies which importers hold warm links to unaffected sources and drafts the engagement pack for trade associations.
Diplomatic & Partnership Agent
G2G INSTRUMENTS
Surfaces which agreements are activatable — rice trade agreements, essential supplies agreements, regionalisation arrangements — and what each would deliver.
Stockpile Agent
DRAWDOWN & ROTATION
Computes coverage days by food group under the scored scenario and models drawdown, rotation and replenishment options.
Local Production Agent
SURGE CAPACITY
Estimates how much additional output local farms can bring on, over what horizon, for eggs, seafood and leafy vegetables.
Food Safety Agent
ASSURANCE GATE
Verifies that any proposed new source meets accreditation, certification and testing requirements. Holds a veto over every substitution.
Public Communication Agent
CONFIDENCE MANAGEMENT
Drafts staged public and trade messaging. Panic buying is a demand-side shock the platform must anticipate, not merely observe.
Authority Boundary
GOVERNING RULE
Agents do not determine risk and do not decide. They act on a risk value produced by the patented kernel and stage options for an accountable officer. Nothing is issued externally without human authorisation.

Tower Processing Pipeline

STAGE 01

Sense

Multilingual ingestion across veterinary, phytosanitary, trade policy, climate, maritime, commodity and administrative sources.

STAGE 02

Verify

Adversarial multi-agent falsification and source reliability scoring before any event enters the national risk register.

STAGE 03

Score

Trigger event compared against geospatial and weather data. Risk value computed with positive and negative components; impact percentile against threshold.

STAGE 04

Reason

Inference and decision engines convert the scored risk into national consequence and a ranked set of pillar actions.

STAGE 05

Act

Operating agents stage the response; the accountable officer authorises. Every action is logged against the kernel value that justified it.

Stage 03 implements the characterising portion of granted Singapore patent SG 10202109991Y, whose specification expressly describes a command-and-control centre for monitoring and controlling food supply and demand operations.
COMMAND › DECISION KERNEL CHAIN

Decision Kernel Chain

The architectural principle that governs the platform: the patented risk determination methodology is the immutable core, and every other capability is a modular layer around it. AI models predict. The LLM understands. The kernel determines. The inference engine reasons. Agents stage. Officers decide. Without this discipline a national deployment degrades into a general-purpose AI assistant with no auditable basis for the numbers it reports.

INPUT · EXTERNAL WORLD
Global Food System Intelligence
Weather and harvest conditions, animal and plant disease notifications, trade policy and export restrictions, geopolitics and chokepoints, commodity and input markets, news and vernacular reporting across every source country.
INPUT · NATIONAL ADMINISTRATIVE
SFA Data Sources & Sensing
Import permits and declarations, the accredited establishment registry, inspection and laboratory results, local farm production returns, stockpile records, wholesale market and fishery port throughput, cold store telemetry.
LAYER 01
AI / LLM Intelligence
LLM and NLP extract structured events from unstructured text in many languages. Predictive models forecast supply, disease propagation, route transit, consumption, price and local production. Both produce signals, never risk verdicts.
LAYER 02 · SG 10202109991Y · IMMUTABLE CORE
Patented Risk Determination Kernel
Trigger event detection and validation · geospatial correlation in real time · weather correlation · risk value with positive and negative scores · impact percentile against threshold · high / intermediate / low classification · supply-side and demand-side attribution · sensor-based alerting.
Authoritative risk sourceGranted 6 Dec 2023Priority 11 Sep 2020Command-and-control centre disclosed in spec
LAYER 03
Inference Engine
Answers what does this mean for Singapore. Propagates a scored event through the national ontology to every affected source country, establishment, food item, importer, route and stockpile line.
LAYER 04
Decision Engine
Answers which pillar to pull, in what order. Generates and prices the option set across Diversify Imports, Global Partnerships, Stockpile and Grow Local, with trade-offs stated.
LAYER 05
Agentic Staging
Accreditation, industry engagement, diplomatic, stockpile, local production, food safety and communication agents assemble the executable response for officer authorisation.
LAYER 06 · SURFACE
National C3 & Officer Decision
Command surface, decision record, escalation to senior management and ministry, and the audit trail that lets a national decision be explained afterwards.

Worked Example — Avian Influenza in a Major Poultry Source

SignalAnimal health watch detects clustered poultry mortality reporting in a producing region of a major source country, ahead of formal notification.
LLM extractionLocal-language reports, veterinary bulletins and state agriculture notices parsed into structured events with region, species, farm count and effective date.
Model predictionDisease propagation model projects the affected zone over 60 days. Supply model projects the share of Singapore's poultry and egg volume originating inside that zone. Price model projects the cost of substitution.
Kernel determinationTrigger event compared against geospatial establishment locations and weather data. Risk value computed, impact percentile tested against threshold. Result: poultry supply risk HIGH; 22% of volume inside the projected zone; egg supply risk INTERMEDIATE.
InferenceUnder regionalisation arrangements, establishments outside the zone remain eligible. Three accredited alternate sources have spare export capacity. Frozen substitution absorbs part of the gap. Local egg production covers a further share.
DecisionPillar sequence: Diversify first (redirect to unaffected zones and accredited alternates), Partnerships second (confirm regionalisation with the source authority), Stockpile third (hold, do not draw), Grow Local (signal egg farms to sustain output).
Agentic stagingAccreditation agent ranks candidate establishments; food safety agent verifies eligibility; industry agent identifies importers with existing links; communication agent drafts a trade circular and a public reassurance line.
Officer decisionAuthorising officer approves the circular and the engagement pack; stockpile drawdown withheld; risk value, evidence and rationale written to the decision record.
This mirrors what Singapore actually did when avian influenza was detected on a Brazilian poultry farm in 2025: imports continued from unaffected areas because regionalisation agreements were already in place. The platform's contribution is not the response — SFA already knows the response. It is compressing detection and quantification so the same response begins weeks earlier and is documented as it happens.

Separation of Concerns

ComponentQuestion it answersWhat it must never doIP status
Predictive AI modelsWhat is likely to happen?Declare a risk level or trigger a responseTRADE SECRET
LLM / NLP layerWhat does this text mean?Score risk, or act as the reasoning authorityCONFIGURATION
Patented risk kernelHow much risk, where, and does it cross threshold?Be bypassed, duplicated or overridden by any other layerGRANTED PATENT
Inference engineWhat does this risk mean for the national supply?Recompute the risk valueFILING CANDIDATE
Decision engineWhich pillar, in what order, at what cost?Select a pillar without pricing the alternativesFILING CANDIDATE
Agentic layerWhat has to be prepared, and by whom?Issue anything externally without officer authorisationFILING CANDIDATE
OfficerWhat will Singapore do?Be presented with a recommendation whose basis cannot be tracedACCOUNTABLE
COMMAND › SFS2 FOUR PILLARS COMMAND

Our Singapore Food Story 2 — Four Pillars Command

SFA's refreshed food resilience strategy, announced in November 2025, rests on four pillars: Diversify Imports, Global Partnerships, Stockpile and Grow Local. Together they give Singapore options and flexibility against different disruptions. The platform's entire purpose is to tell an officer which pillar to pull, when, and at what cost — with the sequencing evidenced rather than asserted.

PILLAR 01 SFA
Diversify Imports
Reduce source concentration and improve agility to source in times of need. Over 180 countries and regions supply food today, up from about 140 two decades ago.
Platform roleSource scoring · substitution readiness
PILLAR 02 SFA
Global Partnerships
Government-to-government collaborations that keep supply flowing during disruption. Rice trade agreements with Cambodia, Thailand and Vietnam; an essential supplies agreement with New Zealand; a joint agri-tech food zone study with Brunei.
Platform roleActivation triggers · instrument register
PILLAR 03 SFA
Stockpile
The buffer during severe disruption, now covering all three key food groups — carbohydrates, protein and fibre — focused on longer shelf-life items including rice, frozen protein and canned vegetables.
Platform roleCoverage clock · drawdown modelling
PILLAR 04 SFA
Grow Local
The one regenerative source. By 2035, local farms aim to supply 20 per cent of fibre and 30 per cent of protein consumed locally; as of 2025 they contributed about 8 per cent and 25 per cent.
Platform roleSurge headroom · productivity tracking

Pillar Readiness Against Threat Class

How much relief each pillar can deliver against each of SFA's three published risk classes, and over what horizon. Prototype scoring. The pattern that matters: diversification is fast but shallow against a region-wide climate event; stockpile is deep but finite; local production is the only line that regenerates.

Pillar Sequencing Rules

Scenario classFirst pillarSecondThirdSequencing logic
Single-source disease outbreakDiversifyPartnershipsGrow LocalRegionalisation preserves most of the source; substitution covers the remainder. Stockpile held in reserve.
Source-country export banPartnershipsDiversifyStockpileDiplomatic exemption is the fastest lever when the constraint is political rather than physical.
Region-wide climate eventStockpilePartnershipsDiversifySubstitution fails when the whole producing region is affected. Buffer first, then agreements, then distant sources.
Route or chokepoint closureDiversifyStockpilePartnershipsReroute and re-source; buffer absorbs the transit gap while longer routes stand up.
Prolonged multi-channel crisisGrow LocalStockpilePartnershipsStockpiles run out. Local production is the only regenerative source in a long crisis.
Perishable, low-substitutability itemGrow LocalDiversifyStockpileLeafy greens cannot be sourced from far away or stockpiled. Local surge is the only real buffer.
Sequencing rules are prototype logic derived from SFA's published description of how the pillars work together, including the treatment of carbohydrates, proteins and vegetables in the Singapore Food Story 2 material.

Platform Coverage of the Mandate

PillarOfficer questionPlatform viewKernel contribution
Diversify ImportsWhich sources are at risk, and which alternates are ready?05 · 09 · 12 · 17Per-source risk value and open-probability
Global PartnershipsWhich agreement should we activate, and when?04 · 28 · 39Threshold crossing as the activation trigger
StockpileHow many days of cover remain under this scenario?22Demand-side risk classification and drawdown modelling
Grow LocalHow much can local farms absorb, and how fast?16 · 18Local production as the negative-score offset in the risk value
Food safetyIs the substitute source actually eligible?26Assurance gate on every substitution
Public confidenceWhat do we say, and when?18 · 28Panic-buying modelled as demand-side risk
COMMAND › SOURCE & CORRIDOR BOARD

Source Country & Corridor Board

Diversification is a portfolio, and a portfolio has to be scored. Every food item carries a ranked list of source countries with live open-probability, accreditation status, transit distribution and substitution cost. When one degrades, the alternate set is already priced and already eligible.

SOURCE COUNTRIES SFA 2025
180+
Supplying food to Singapore
VEGETABLE SOURCES SFA
100+
Countries and regions
EU SOURCES FOR MEAT & EGG SFA 2026
18
Member states, up from 15
SOURCES IN ALERT ILLUSTRATIVE
7
Amber or red this cycle

Source Concentration by Food Item

Effective source concentration against count of accredited or approved sources. Items in the upper-left quadrant — few sources and high concentration — are where a single event becomes a national issue. Prototype scoring.

Recent Source Widening

PeriodItemNew sources approvedEffect on the portfolio
2024 SFAPorkPortugalAdds a European chilled and frozen option
2024 SFABeefBrunei, PolandAdds one regional and one European source
2024 SFAPoultryTürkiyeWidens the non-regional poultry set
2025 SFAShell eggsLithuaniaReduces regional egg concentration
2025 SFAProcessed eggsBrazil, Lithuania, ParaguayAdds shelf-stable protein optionality
2025 SFAPoultryGreece, Lithuania, ParaguayThree additional sources in one year
2025 SFAPorkGreece, ParaguayBroadens beyond the traditional set
2025 SFABeefGreece, Paraguay, Republic of KoreaAdds a North Asian option
Apr 2026 SFAMeat and eggGreece, Lithuania, LatviaBrings approved EU sources to 18 member states
Aug 2026 SFAPoultryBrunei DarussalamAdds a close regional poultry source
This table is the diversification pillar made visible. The platform's contribution is to rank the next accreditation candidates by the risk they would retire, rather than by the order requests arrive.

Substitution Readiness

If this is disruptedPrimary substitutionForm change requiredActivation leadCost deltaReadiness
Regional fresh poultryChilled and frozen from accredited alternatesFresh to chilled or frozen2–6 weeks+12%PROVEN 2022
Regional live porkChilled and frozen from 20+ sources; alternate live sourceFresh to chilled or frozenDays to weeks+9%PROVEN 2023
Non-basmati rice, single originOther varieties and origins; trade agreement activationVariety change4–10 weeks+18%PROVEN 2023
Regional leafy vegetablesAlternate regional sources plus local surgeOrigin and quality change1–3 weeks+24%PARTIAL
Regional shell eggsAccredited alternates plus local productionOrigin change3–8 weeks+16%PARTIAL
Mekong-wide rice failureDistant origins; stockpile drawVariety and price10–20 weeks+40%CONSTRAINED
Multi-route closureAir freight for perishables; longer sea routesModality change2–5 weeks+55%CONSTRAINED
Readiness marked PROVEN corresponds to a documented Singapore response: the 2022 Malaysian chicken export ban, the 2023 Pulau Bulan African Swine Fever detection, and the 2023 Indian non-basmati rice ban. Cost deltas are prototype estimates.
COMMAND › SUPPLY CHAIN COVERAGE

Food Supply Chain Coverage

Thirteen nodes from the source farm to the consumer. The kernel sits across all of them rather than at a single point, because the events that cause a national shortage propagate between nodes — a fertiliser shock becomes a yield shortfall, a yield shortfall becomes an export restriction, an export restriction becomes an empty shelf and then a panic-buying event that is itself a second shock.

Source Farm
YIELD · DISEASE · WEATHER
Source Processing
ACCREDITED ESTABLISHMENT
Export Certification
VETERINARY · PHYTO
Port of Loading
CAPACITY · CONGESTION
Sea / Air Route
CHOKEPOINTS · FREIGHT
Port of Entry
DECLARATION · PERMIT
Inspection
RISK-BASED CHECKS
Laboratory Testing
NCFS · TURNAROUND
Cold Store
TEMP · POWER · SPACE
Wholesale
PASIR PANJANG · JURONG
Distribution
TRADE · F&B · CATERING
Retail
SUPERMARKET · WET MARKET
Consumer
PRICE · CONFIDENCE

Instrumentation Depth Against Risk Materiality

Where the risk is high and the sensing is thin is where phase one investment goes. Prototype scoring against a notional current-state baseline.

Node Detail

NodeMonitored signalsDominant trigger eventsKernel outputState
Source farmWeather, harvest condition, disease notifications, herd and flock statusDrought, flood, outbreak, feed shortageSource exposure scoreAMBER
Source processingAccreditation status, audit findings, capacitySuspension, delisting, capacity lossEstablishment eligibilityGREEN
Export certificationHealth certificate issuance, regionalisation zonesCertificate suspension, zone redrawCorridor open-probabilityGREEN
Port of loadingBerth availability, congestion, reefer plug capacityCongestion, strike, closureDeparture reliabilityAMBER
Sea / air routeAIS, chokepoint transits, war-risk premia, air cargo ratesChokepoint closure, rerouting, capacity lossTransit distributionRED
Port of entryPermit declarations, arrival volumes, clearance timesVolume shortfall, clearance backlogArrival adequacyGREEN
InspectionInspection rates, detentions, rejectionsRejection cluster, capacity constraintAssurance loadGREEN
Laboratory testingSample volumes, turnaround, non-compliance rateBacklog, emerging contaminantTesting headroomAMBER
Cold storeUtilisation, temperature, power continuityCapacity ceiling, excursion, outageBuffer integrityAMBER
WholesaleThroughput, arrivals, price formationVolume drop, price spikeMarket signalGREEN
DistributionTrade and catering demand, establishment activityChannel demand shiftChannel balanceGREEN
RetailAvailability, price, form substitutionStockout, price pass-throughAvailability indexGREEN
ConsumerSearch interest, sentiment, purchase surgesPanic buying, confidence lossDemand-side riskAMBER
The consumer node is a genuine risk node, not a reporting endpoint. SFA's own guidance asks the public to stay calm and avoid panic buying during disruptions — which means panic buying is a modelled demand-side shock, and the patent's supply-side versus demand-side classification applies directly to it.
INTELLIGENCE › SIGNAL SCANNER

Signal Scanner

Weak-signal detection is the whole value of the system. Every disruption in the past disruption history had a public, machine-readable precursor. The question for a national platform is never whether the signal existed — it is whether anyone was reading it in the right language, on the right day, with the right exposure map beside it.

SIGNALS INGESTED / 24H
9,640
Across 12 source classes
SURVIVED VERIFICATION
284
2.9% pass rate
ABOVE THRESHOLD
23
Entered the national risk register
PLAYBOOKS ARMED
5
Staged for officer decision

Live Signal Feed

DetectedSignalSource classFood groupConfidenceRisk
T-0hHighly pathogenic avian influenza notifications clustering across multiple approved sources
Northern hemisphere season approaching; overlapping poultry and egg exposure
Animal healthPoultry, eggs0.91HIGH
T-5hMekong basin hydrology below seasonal normal
Rice yield model flags a region-wide rather than country-specific signal
ClimateRice0.77MED
T-11hFertiliser freight and price dislocation persisting
Traced to Strait of Hormuz disruption; two to three season lag to yields
Input marketAll crops0.88MED
T-1dExport restriction language appearing in a source-country ministry statement
Pre-announcement pattern matches the 2022 and 2023 precedents
Trade policyPoultry0.64MED
T-2dProlonged heavy rainfall in regional vegetable-growing highlands
Quality degradation precedes volume shortfall by roughly ten days
ClimateLeafy vegetables0.82MED
T-3dAfrican swine fever activity in the regional live pig chain
Live pig price surge is the confirming market signal
Animal healthPork0.73MED
T-4dAir cargo capacity tightening on perishable lanes
Reduces the fallback modality for chilled substitution
LogisticsChilled meat, seafood0.69MED
T-6dLocal egg farm operational disruption
Assessed as localised; monitored for systemic spread
Local productionEggs0.95LOW
T-8dConsumer search interest in a staple item rising above baseline
Early panic-buying precursor; demand-side classification
Public sentimentRice, eggs0.58MED
Illustrative prototype feed. Confidence is the post-verification posterior, not raw source confidence.

Signal Class Yield

INTELLIGENCE › SOURCE ESTABLISHMENT INTELLIGENCE

Source Establishment & Local Farm Intelligence

Diversification is counted in countries but delivered by establishments. A source country with one accredited plant is not diversified. This view scores the accredited overseas estate and the local farm base together, because in a disruption they are two halves of the same substitution calculation.

ACCREDITED SOURCE COUNTRIES SFA 2024
42
Approved for higher-risk items: livestock, meat, eggs
LAND-BASED FARMS SFA 2024
153
Down from 156 in 2023
SEA-BASED FARMS SFA 2024
72
Down from 98 in 2023
VEG FARM PRODUCTIVITY SFA 2025
253t/ha
Up about 10% from 231.4

Local Production Trajectory Against the 2035 Targets

Local contribution to fibre and protein consumption against the 2035 targets of 20 per cent and 30 per cent. Historical points are SFA-published; the forward path is a prototype projection, not a forecast endorsed by SFA.

Fibre — 8% today, 20% by 2035 SFA

Fresh leafy and fruited vegetables, beansprouts and mushrooms. The widest gap to target, and the category where local production matters most because leafy greens are the least substitutable from distant sources.

Productivity +10% in 2025Controlled environment shift

Protein — 25% today, 30% by 2035 SFA

Eggs and seafood. Closest to target. Seafood farm productivity rose about 27 per cent in 2025, from 40.7 to 51.5 tonnes per hectare per year; egg output saw a slight decline attributed to localised operational issues.

Seafood productivity +27%National Broodstock Centre

The productivity story

Farm counts have fallen while output per hectare has risen. That is the signal the platform should track — resilience comes from productive capacity under controlled conditions, not from the number of licences issued.

ACT Fund 2Hatchery Development & Recognition

Establishment Exposure Board

EstateScopeConcentration riskMonitored signalsSurge headroomExposure
Accredited meat establishmentsOverseas, higher-risk itemsUneven across countriesAudit findings, suspensions, disease zonesSource-dependentMEDIUM
Accredited egg establishmentsOverseas plus localHistorically concentratedHPAI zones, flock status, certificationLimited near termHIGH
Vegetable supply base100+ countries and regionsLeafy greens regionally concentratedRainfall, quality, harvest volumesLocal surge, weeksMEDIUM
Rice supply baseMekong-weighted plus alternatesRegion-level correlationHydrology, export policy, stocksAgreements + stockpileMEDIUM
Local vegetable farmsLand-based, controlled environmentSmall base, rising productivityYield, productivity, farm countBuffer, not replacementMEDIUM
Local egg farmsLand-basedFew operatorsFlock health, output, operationsMeaningful share of proteinMEDIUM
Local fish farmsSea-based and land-basedConsolidating estateWater quality, algal blooms, harvestImproving with technologyLOW
Closed containment aquaculture and controlled-environment growing reduce exposure to environmental hazards such as algal blooms — which is precisely why productivity per hectare, not licence count, is the resilience metric that belongs on a command board.
INTELLIGENCE › DISEASE & REGIONALISATION

Transboundary Disease & Regionalisation

SFA identifies disease outbreaks as one of three principal risks to Singapore's food supply, and notes that unlike crops, livestock take longer to breed and grow — a single incursion can disrupt meat supplies for months or years. Regionalisation is the instrument that keeps a source open when only part of it is affected, and it only works if the affected zone can be drawn quickly and defended with evidence.

ACTIVE DISEASE WATCHES
9
Across source jurisdictions
HIGHEST BAN HAZARD
0.41
180-day, poultry and egg sources
VOLUME IN WATCHED ZONES
22%
Of poultry volume, projected zone
MEAN DETECTION LEAD
21d
Ahead of formal notification

Volume Exposure Against Disease Risk by Source Group

Share of national volume held by each source group, plotted against that group's disease exposure score. Concentration in a high-exposure group is the condition under which a single outbreak becomes a national issue. Prototype scoring.

Determination Sequence

Where is the event?Geospatial zone resolved from official notification, veterinary reporting and vernacular sources, ahead of and then reconciled with formal channels.
Which establishments are inside?Accredited establishment coordinates intersected against the zone and the projected propagation front.
What share of national volume?Exposed volume as a percentage of the item's national supply, weighted by whether the affected form is substitutable.
Is regionalisation available?Whether an arrangement exists with the source authority, and whether the unaffected zone can be evidenced to Singapore's standard.
Probability and impact?Ban or suspension hazard over 180 days multiplied by the value and criticality of the exposed volume.
Threshold crossed?Impact percentile tested against a threshold modulated by stockpile coverage days for that food group — the less cover, the earlier the same raw risk crosses.
ClassificationHigh / intermediate / low per the granted claim, then supply-side or demand-side, then per source and per food item.

Disease Watch Register

DiseaseItems at riskLeading indicatorRegionalisation availableBan hazard 180dClass
Highly pathogenic avian influenzaPoultry, eggsWild bird surveillance, mortality clusters, seasonalityYes, with several sources0.41HIGH
African swine feverPork, live pigsRegional case reporting, live pig price surgesPartial0.28INTERMEDIATE
Foot and mouth diseaseBeef, mutton, live ruminantsClinical description clusters, movement controlsYes, zone-based0.22INTERMEDIATE
Lumpy skin diseaseBeefVector season, regional notificationsPartial0.17INTERMEDIATE
Aquatic disease and algal bloomSeafood, local aquacultureWater quality, temperature, bloom indicesNot applicable0.11LOW
Plant pest and phytosanitary eventsVegetables, fruitIPPC notifications, interception patternsYes, area-based0.09LOW
Ban hazard figures are prototype model outputs. The register is scored, not merely listed: every row feeds the kernel as trigger event data and contributes to the food group assurance shown on the executive board.
INTELLIGENCE › CONSUMPTION & DEMAND

Consumption & Demand Intelligence

The patent distinguishes supply-side from demand-side risk. For a national platform the demand side is not a sales forecast — it is national consumption by food group, the substitution behaviour that occurs under stress, and the confidence dynamics that can turn an adequate supply into a visible shortage overnight.

NATIONAL SUPPLY ADEQUACY ILLUSTRATIVE
96%
Across all food groups, current conditions
GROUPS BELOW COMFORT
2
Poultry and eggs; leafy vegetables
SUBSTITUTION HEADROOM
High
Frozen and chilled forms available at scale
CONFIDENCE INDEX
Stable
No panic-buying precursor above threshold

Consumption Against Assured Supply by Food Group

Assured supply and at-risk share by food group under current scored conditions. Prototype figures. Note the shape: total volume is comfortable almost always; the risk concentrates in specific groups with low substitutability.

Demand-Side Register

Food groupConsumption characterSubstitution behaviour under stressConfidence sensitivityFailure consequenceState
RiceStaple, inelasticVariety switching, reluctantlyVery highVisible staple shortage; strong public reactionAMBER
Poultry & eggsDaily protein, price-visibleFresh to chilled to frozen; proven in 2022HighPrice pass-through, hawker cost pressureRED
PorkCultural staple in segmentsFresh to chilled and frozen; proven in 2023Medium-highSegment-specific dissatisfactionAMBER
Beef & muttonDiscretionary and festiveOrigin and grade switching, easyLowPrice effect onlyGREEN
SeafoodDaily and festiveSpecies and form switchingMediumSpecies availability shiftsGREEN
Leafy vegetablesDaily, highly perishablePoor; quality degrades before volume fallsHighVisible quality decline at wet marketsAMBER
FruitDiscretionaryWide origin substitutionLowPrice and variety effectGREEN
Catered & institutional foodBulk, contract-drivenMenu substitution at short noticeLowContract pressure; food safety load risesGREEN

Confidence and Panic-Buying Model

Why it belongs in the kernelA surge in household purchasing is a demand-side trigger event with a measurable geospatial and temporal footprint. It meets the claim's definition as squarely as a drought does, and it can be scored on the same scale.
Leading indicatorsSearch interest for staple items, social media velocity around a supply story, wholesale order acceleration, retail replenishment frequency, and the gap between reported and actual availability.
Why lead time matters most herePanic buying is the one shock where the intervention — clear, early, specific public communication — is cheap and fully within the agency's control. Every day of warning is disproportionately valuable.
The governing constraintA false positive here is expensive in public trust. This is the single strongest argument for the adversarial verification layer sitting between sensing and scoring.
Documented precedentWhen Malaysia halted chicken exports in 2022, the mitigation depended on hawkers, retailers and consumers adapting to alternative sources and frozen forms. Consumer behaviour was not a side effect of the response — it was part of it.
INTELLIGENCE › PAST DISRUPTION HISTORY

Past Disruption History

Six documented disruptions, five of them Singapore's own. Each is replayed through the tower to calibrate thresholds and to establish how much earlier the same correct response could have started. The point is never that the response was wrong — Singapore's responses were right. The point is when they began.

Achievable Warning Lead Time by Past Disruption

Weeks of actionable warning available from public precursor signals, measured to the date the disruption first bound national supply. Prototype assessment.

Malaysian Chicken Export Ban, June 2022

The canonical Singapore case: a neighbouring country's domestic shortage became Singapore's supply problem, and the mitigation was a whole-of-nation effort rather than a single lever.

EARLY 2022
Feed cost escalation from the Russia–Ukraine conflict compresses regional producer margins. The upstream signal precedes the trade action by months.
MAY 2022
Domestic price and supply pressure builds in the source country. Price control and subsidy measures appear — the standard pre-ban policy sequence.
JUN 2022
Malaysia bans chicken exports to address its own domestic shortage. Singapore loses a key source of fresh chicken.
RESPONSE
SFA accelerated accreditation of new sources. Businesses expedited chilled chicken imports from other sources including Australia and Thailand. Hawkers and stallholders switched with support from the Meat Traders' Association. Consumers adapted to frozen poultry and other meats.
OUTCOME
Impact mitigated quickly because every actor did their part. Singapore now imports chicken from more sources than before.
Platform contribution: the feed-cost-to-export-ban pathway is a repeatable, machine-detectable sequence. The Trade Policy agent watches for the same policy language in every source country, and the accreditation agent pre-ranks candidates before the ban is announced rather than after.
LEAD TIME AVAILABLE: 6–10 WEEKS Poultry Class: Trade policy, feed-driven

African Swine Fever, Pulau Bulan, April 2023

A major fresh pork source was cut off overnight — and supply did not collapse, because alternatives had been built years earlier.

EventAfrican Swine Fever detected in a consignment of pigs from Pulau Bulan, Indonesia. Singapore halted imports to ensure food safety.
Immediate effectA major source of fresh pork removed overnight.
ResponseSFA worked with industry partners including the Meat Traders' Association to increase supplies of chilled and frozen pork. Consumers could choose from more than 20 sources including Australia, Brazil, Denmark and New Zealand.
The decisive detailLive pig imports from Sarawak — approved years earlier — were quickly ramped up. The option existed because someone had built it before it was needed.
Platform lessonThe value of an accreditation is not realised on the day it is granted. It is realised on the day a different source fails. That argues for ranking accreditation candidates by the risk they would retire, which is exactly what the source board does.
LEAD TIME AVAILABLE: 2–5 WEEKS Pork Class: Disease

India's Non-Basmati Rice Export Ban, July 2023

The world's biggest rice exporter banned exports of non-basmati rice. Singapore's response used two pillars at once.

JUL 2023
India bans non-basmati rice exports.
DIPLOMATIC
Singapore engaged Indian authorities and obtained an exemption from the ban to maintain supply flows. The Global Partnerships pillar, used as a first rather than a last resort.
COMMERCIAL
Importers increased imports of different rice varieties from multiple different sources. The Diversify pillar working in parallel.
Platform contribution: export restriction announcements are rarely the first signal. Domestic price pressure, procurement policy shifts and stock position statements in the exporting country precede them. The decision engine's contribution here is sequencing — it identifies that a diplomatic instrument outranks commercial substitution when the constraint is political.
LEAD TIME AVAILABLE: 3–8 WEEKS Rice Class: Trade policy

Avian Influenza, Brazil, May 2025

Bird flu was detected on a poultry farm in Brazil. Singapore had no poultry from the affected farm, and was still able to import chicken from unaffected areas because regionalisation agreements were already in place.

The cleanest demonstration of why zone-level rather than country-level risk determination matters. The kernel's geospatial correlation is the technical basis for exactly this distinction.
CONTAINED Poultry

Heavy Rainfall, Melaka and Cameron Highlands, 2022

Prolonged heavy rain depleted crop yield from many vegetable farms, reducing supply to Singapore. Vegetables were often of poorer quality and spoiled more quickly. Importers sourced from alternative countries; locally grown vegetables cushioned the shortfall.

Quality degrades before volume falls. A platform watching only tonnage would have seen this late; a platform watching rainfall and quality indices sees it first.
MITIGATED Leafy vegetables

Chokepoint Disruptions, 2021 and 2026

In March 2021 a 400-metre container ship grounded in the Suez Canal, blocking it for six days and causing a backlog of over 400 vessels and months of schedule disruption. In 2026, conflict in the Middle East led to a blockade of the Strait of Hormuz, disrupting not only energy but fertiliser flows — with higher logistics costs, lower agricultural yields and potential food supply disruption downstream.

The Hormuz case is the clearest argument for upstream input monitoring: fertiliser is where a distant conflict enters the food system, and the lag is measured in growing seasons.
ACTIVE All crops

What the Precedent Set Establishes

Every disruption had a precursorFeed costs before the chicken ban. Regional disease activity before the pork halt. Domestic price pressure before the rice ban. Rainfall before the vegetable shortfall. Conflict before the fertiliser dislocation.
The right response was knownIn every case Singapore's response was correct and effective. This is not a competence gap.
The gap is time, and evidenceResponses began after the disruption bound supply. Detection compression converts a reactive mitigation into a pre-positioned one, and produces a defensible record of why each decision was taken.
The second-order effects are the hard partThe 2022 feed shortage produced export bans months and years later. Chains of consequence across time are what the inference engine exists to trace, and what no dashboard does.
INTELLIGENCE › PREDICTIVE MODEL BANK

Predictive Model Bank

Six forecasting models feed the kernel. Three are proprietary and constitute the analytical moat; three are competent implementations of established methods. None declares a risk level — they emit signals that the patented kernel converts into a risk value.

MODEL 01 · PROPRIETARY CORE

Disease & Trade Restriction Propagation

Given an outbreak or policy signal at a location, predicts the affected zone over time and the probability and timing of an export restriction or import suspension affecting Singapore's supply of that item.

CORPUS
WOAH notificationsHistoric ban registerRegionalisation precedentsTrade policy language
MOAT Output: 180-day restriction hazard curve
MODEL 02 · PROPRIETARY CORE

Corridor Transit & Arrival Distribution

Predicts arrival time and its variance per origin-item-modality combination, including chokepoint routing, port congestion, inspection and testing dwell, and cold chain constraints.

CORPUS
AIS transitsArrival recordsClearance timesAir cargo rates
MOAT Output: P50 / P90 arrival and reliability
MODEL 03 · PROPRIETARY CORE

Substitution Feasibility

For any disrupted item, computes which alternate sources are simultaneously accredited, available, affordable and arrivable within the window. Answers whether the diversification portfolio is real or nominal.

CORPUS
Accreditation registryExport capacityHistoric substitution episodes
MOAT Output: ranked feasible substitution set
MODEL 04

Consumption & Confidence Model

Forecasts national consumption by food group and models demand-side surge behaviour, including panic-buying precursors and form substitution under stress.

STANDARD METHOD
MODEL 05

Climate, Harvest & Local Production

Projects source-region yields from hydrology, ENSO state and crop condition indices, and projects local farm output and surge headroom for eggs, seafood and vegetables.

STANDARD METHOD
MODEL 06

Input, Price & Cost Model

Projects feed, fertiliser, energy and freight costs and their transmission into food prices and yields, with the multi-season lag that makes upstream inputs a leading rather than coincident indicator.

STANDARD METHOD

National Supply Adequacy Forecast — Distribution

Illustrative Monte Carlo over source open-probability, transit variance and local production for a single at-risk food group. The comfort line is the operative constraint — the distribution mass falling below it is the number that belongs in front of senior management.

Model Governance

ModelRefreshBacktest windowHuman overrideFailure mode watched
Disease & restriction propagationDaily2000–2026Required above thresholdPolitical restrictions without sanitary basis
Corridor transitDailyRolling 36 monthsAdvisoryRegime change after chokepoint closure
Substitution feasibilityWeeklyDocumented episodesRequired at activationNominal accreditation mistaken for live capacity
Consumption & confidenceDailyRolling 5 yearsRequired before any public messageReflexivity — the warning causing the surge
Climate & local productionDailyRolling 10 yearsAdvisoryRegional correlation understated
Input, price & costDaily2015–2026AdvisoryLag misspecification across seasons
The consumption and confidence model carries the strictest override rule in the bank. A model that can influence the behaviour it predicts must never drive a public communication without an accountable human decision.
INTELLIGENCE › NATIONAL RISK MATRIX

National Risk Mapping Matrix

Likelihood against impact for the risks that bind Singapore's food supply. Each is anchored to a documented event or to SFA's published risk taxonomy rather than to judgement alone.

LIKELIHOOD (X) × IMPACT (Y) — CELL VALUES ARE RISK IDENTIFIERS
SCALE

National Risk Register

IDRiskEvidence anchorLikelihoodImpactDetectable leadRating
R1Transboundary animal disease in a major sourceASF Pulau Bulan 2023; HPAI Brazil 2025; ASF Malaysia 2025HighSevere3–5 wkCRITICAL
R2Source-country export restrictionMalaysia chicken 2022; India non-basmati rice 2023HighSevere6–10 wkCRITICAL
R3Region-wide climate event affecting a stapleSFA: Mekong rice concentration; harvests take months to years to recoverMed-HighSevere8–20 wkCRITICAL
R4Chokepoint or route closureSuez 2021; Hormuz 2026MediumMajor2–4 wkHIGH
R5Upstream input shock — feed, fertiliser, energyRussia–Ukraine 2022 feed shortage; Hormuz fertiliser 2026HighMajor1–3 seasonsHIGH
R6Perishable regional supply failureMelaka and Cameron Highlands rainfall 2022HighModerate1–3 wkELEVATED
R7Multi-channel prolonged crisisCOVID-19 border restrictions across sea and airUnlikelySevereVariableHIGH
R8Panic buying and confidence lossSFA public guidance to avoid panic buying during disruptionsMediumMajorDaysHIGH
R9Local production capacity erosionFarm counts declined 2023 to 2024; fibre at 8% against a 20% targetMediumModerateQuartersELEVATED
R10Food safety incident in an imported consignmentOngoing recalls and adulteration findingsLikelyMinorReal timeELEVATED
R11Cold chain or storage failureCold store capacity, power and temperature exposureUnlikelyMajorReal timeELEVATED
R12Accreditation or certification lapse at sourceEstablishment-level suspension exposureMed-HighMinor4 wkELEVATED
OPERATIONS › STOCKPILE & COVERAGE CLOCK

Stockpile & Coverage Clock

SFA describes the stockpile as the buffer during severe disruptions, now expanded to cover all three key food groups — carbohydrates, protein and fibre — focused on longer shelf-life items including rice, frozen protein and canned vegetables. For a national platform the stockpile is not an inventory line. It is the clock that determines how much time every other pillar has to work.

COVERAGE CLOCK — DAYS OF NATIONAL SUPPLY UNDER THE CURRENT SCORED SCENARIO ILLUSTRATIVE
72
CARBOHYDRATES
54
PROTEIN
31
FIBRE
BINDING MINIMUM
Figures are prototype placeholders chosen to demonstrate the mechanism. Actual national stockpile levels are not public and are not represented here. The binding minimum is the operative number: national resilience is set by the shortest clock, not the average.

Drawdown Under Scenario

Modelled coverage remaining over time under a severe disruption, with and without the other three pillars activated. The gap between the two lines is the value of early pillar sequencing — every week the diversification and partnership pillars are pulled earlier is a week of stockpile preserved.

Threshold Modulation by Coverage

The patent allows the risk threshold to be preset or dynamically set. In a national deployment the natural dynamic variable is coverage days for the affected food group. Identical raw risk should cross threshold earlier when the buffer behind it is thinner.

Coverage remainingThreshold posturePractical effectOfficer implication
> 90 daysRELAXEDOnly high-confidence, high-impact events escalateMonitor; no pillar activation
60–90 daysSTANDARDNormal scoring; substitution options refreshedVerify alternate sources are live
30–60 daysELEVATEDMedium-confidence events escalate; accreditation candidates rankedPre-position diversification and partnerships
14–30 daysTIGHTWeak signals escalate; local surge modelled continuouslyStage stockpile drawdown options
< 14 daysCRITICALAll signals escalate; demand-side management engagedWhole-of-government escalation
This is the strongest of the fresh filing candidates: a risk threshold modulated by the remaining buffer of the specific resource at risk. It is absent from the 2020 disclosure and is the mechanism that makes the same kernel behave correctly in both a calm year and a crisis.

Stockpile Register

GroupRepresentative items SFAShelf life characterRotation burdenSubstitution roleState
CarbohydratesRiceLongLowPrimary buffer; supports the staple with the least substitutabilityGREEN
ProteinFrozen proteinMedium-long, energy-dependentMediumCovers the group with the highest disease exposureAMBER
FibreCanned vegetablesLongLowCovers the group with the shortest fresh shelf lifeAMBER
Held with industryWarehouse partnersVariesSharedExtends buffer without full public carrying costGREEN
SFA works closely with industry to maintain the stockpiles, and national stockpile warehousing is operated with commercial partners. Cold-stored frozen protein carries an energy dependency that the platform tracks alongside volume — a stockpile that cannot stay frozen is not a stockpile.
INTELLIGENCE › IMPORT DATA EXPLORER

Import Data Explorer

What Singapore imports, how much, and from where — by food item and by source country. This is the base layer the kernel scores against: a trigger event is only a national risk once it is intersected with the volume and the share it puts at stake. Figures below are published import data; the risk metrics derived from them are the platform's own.

CHICKEN IMPORTED, 2021 CNA / SFA
214,400t
Brazil 48% · Malaysia 34% · USA 8%
VEGETABLES IMPORTED, 2021 CNA / SFA
567,900t
Malaysia 42% · China 27% · Australia 5%
EGGS IMPORTED, 2021 CNA / SFA
2,138m pcs
Malaysia 52% · local 30%
HIGHEST SINGLE-SOURCE SHARE
90%
Mutton from Australia

Import Volume by Food Item, 2021

Volumes as reported for 2021. Vegetables and fruit dominate by tonnage; the items that dominate the risk register are not the same ones, which is the reason concentration and substitutability have to be scored separately from volume.

Source Country Shares by Item, 2021 CNA / SFA

Stacked share of each item by named source country, with the remainder grouped as other countries and regions. The height of the first band is the item's single-source exposure.

Item-Level Import Table

Food itemVolume, 2021Top sourceSecondThirdTop-2 shareConcentration
Chicken214,400 tBrazil 48%Malaysia 34%USA 8%82%HIGH
Shell eggs2,138 m piecesMalaysia 52%Local 30%Others 10%82%HIGH
Pork128,100 tBrazil 39%Indonesia 15%Australia 5%54%MEDIUM
Beef39,400 tBrazil 58%Australia 22%USA 9%80%HIGH
Mutton / lamb15,000 tAustralia 90%New Zealand 7%Others 2%97%CRITICAL
Seafood133,400 tMalaysia 25%Indonesia 17%Vietnam 15%42%LOW
Vegetables567,900 tMalaysia 42%China 27%Australia 5%69%MEDIUM
Fruit433,100 tMalaysia 40%China 13%South Africa 8%53%MEDIUM
Volumes and shares as published for 2021; percentages are rounded at source and do not always sum to 100. Concentration ratings are the platform's own scoring of top-2 share against substitutability, not an SFA classification. Later SFA source approvals — Portugal, Poland, Türkiye, Brunei, Greece, Lithuania, Latvia, Paraguay, Republic of Korea and others — have since widened several of these sets.

Correlated Country Exposure — the Risk a Per-Item View Misses

Scoring each item separately understates the real exposure. A single country event does not respect item boundaries. Two countries sit across most of Singapore's protein and fresh produce at once.

Source countryItems where it is a major sourceSimultaneous exposureEvent type that would bind all of themCorrelation risk
MalaysiaEggs 52% · Vegetables 42% · Fruit 40% · Seafood 25% · significant meat4+ itemsDomestic shortage policy, flooding across growing regions, feed cost shock, land border frictionCRITICAL
BrazilBeef 58% · Chicken 48% · Pork 39%3 proteinsAvian influenza, sanitary suspension, port or route disruption on a long sea legCRITICAL
AustraliaMutton 90% · Beef 22% · Vegetables 5% · Pork 5%4 items, one dominantDrought, biosecurity event, export policy changeELEVATED
ChinaVegetables 27% · Fruit 13%2 itemsRegional climate event, phytosanitary action, freight disruptionELEVATED
IndonesiaPork 15% · Seafood 17%2 itemsAnimal disease, proximity effects, export licensingELEVATED
Vietnam / regionalSeafood 15% · rice2 itemsMekong hydrology, export restrictionELEVATED
This is the analytical case for the patented method in a national setting. The claim requires trigger event data to be compared against geospatial data in real time and a risk value computed across the supply chain. Applied here, one Malaysian rainfall event scores against eggs, vegetables, fruit and seafood simultaneously — which is precisely what happened in 2022, and precisely what a per-item report does not show.

Diversification Working — the Egg Case

Malaysia accounted for 72 per cent of Singapore's egg imports in 2019; by 2021 that had fallen to 52 per cent, while the share of locally produced eggs rose from 26 per cent to 30 per cent. Singapore received its first batch of eggs from Poland in June 2020, and has since added sources including Lithuania. Diversification is not a slogan here — it is a measurable, dated reduction in single-source dependence on the most concentrated staple protein.

The platform's contribution is to run this calculation continuously for every item, and to rank the next accreditation candidate by how much concentration it would retire per unit of effort.

Aggregate Supplier Structure

TOP AGRICULTURAL SUPPLIER, 2022

Malaysia — 21%

Of Singapore's agricultural imports by value, followed by the European Union at 16 per cent and the United States at 9 per cent. Breadth at the country-count level coexists with real weight at the top.

SOURCE NETWORK

187 countries and regions

Up from about 140 in 2004. Over 90 per cent of Singapore's food comes through this network. Around 40 to 42 countries and regions are approved to export the higher-risk items — livestock, meat and eggs.

DIVERSIFICATION IN PRACTICE

Eggs from Poland, oranges from Egypt, milk powder from Uruguay, shrimp from Saudi Arabia

Sources opened specifically to broaden supply. Each one is an option that costs little to hold and matters enormously on the day a primary source closes.

From Import Data to a Scored Risk

StepInputWorked example — a sanitary suspension in Brazil
1 · Trigger eventDisease notification, geolocatedAvian influenza confirmed in a producing state
2 · Geospatial intersectionAccredited establishment coordinatesWhich of Singapore's approved Brazilian establishments fall inside the zone
3 · Volume at stakeImport data by item and sourceChicken is 48% Brazil-sourced on a 214,400 tonne base; the exposed share is the zone's share of that 48%
4 · SubstitutabilityAlternate accredited sourcesMalaysia at 34% and the USA at 8% are the near substitutes; form substitution to frozen widens the set further
5 · Correlation checkCountry exposure across itemsBrazil also supplies 58% of beef and 39% of pork — a country-level event scores against three proteins, not one
6 · Risk valuePositive and negative scoresNegative: volume at risk. Positive: regionalisation keeps unaffected zones open, and alternates hold spare capacity
7 · Threshold testImpact percentile vs thresholdThreshold modulated by remaining coverage days for protein; crosses earlier when the buffer is thin
8 · ClassificationHigh / intermediate / low; supply or demand sideSupply-side, HIGH for chicken, INTERMEDIATE for beef and pork
Every step above uses a claim element of SG 10202109991Y in its original form. Only the inputs have changed — from crops and farms to food items and source countries.
OPERATIONS › INPUT & UPSTREAM RISK

Input & Upstream Risk

Food shocks rarely begin as food shocks. They begin as feed, fertiliser, energy and freight. SFA's own account is explicit: the 2022 Russia–Ukraine conflict led to a shortage of animal feed, which affected protein production in some countries and contributed to export bans in the months and years that followed. This is the layer that gives the platform its longest lead times — measured in seasons, not weeks.

INPUT COST INDEX ILLUSTRATIVE
121
Rebased 100 = Jan 2026
FERTILISER DISLOCATION
Active
Traced to Strait of Hormuz disruption, 2026
TRANSMISSION LAG
2–3
Growing seasons to yield effect
EARLIEST WARNING CHANNEL
Upstream
Longest lead time of any signal class

Upstream Transmission Chain

Indexed input costs against the modelled downstream effect on protein availability and food price, with the lag applied. The gap between the two lines is the warning window. Prototype series calibrated on the documented 2022 sequence.

Documented Transmission Pathways

Upstream shockFirst-order effectSecond-order effectReaches Singapore asObserved lag
Russia–Ukraine conflict, 2022Global wheat and corn export halt; Russia and Ukraine account for roughly 25–30% of world wheat and 15% of corn exportsChicken feed costs rise sharply for regional producersHigher egg prices in Singapore and Malaysia; contributing pressure toward export restrictionsMonths
Feed cost escalation, 2022Regional poultry producer margins compressDomestic price and supply pressure in a source countryMalaysian chicken export ban, June 20223–4 months
Strait of Hormuz disruption, 2026Energy and fertiliser flows disruptedHigher logistics cost, lower agricultural yieldsPotential food supply disruption across crop categories2–3 seasons
Energy price shocksCold chain and controlled-environment farming costs riseLocal production economics and stockpile carrying costs deteriorateReduced local buffer, higher frozen stockpile costQuarters
Every row is a chain of consequence across time. Tracing these chains is what the inference engine exists for; a dashboard shows the first column and stops.

Input Exposure Register

InputWhy it matters to SingaporeWatched indicatorSubstitutabilityState
Animal feed — corn, soymeal, wheatSets the cost of regional egg, poultry and pork production, and therefore the probability of export restrictionForward curves, FX, freight, producer marginMediumAMBER
FertiliserSets crop yields one to three seasons out across every vegetable, fruit and grain sourcePrice, freight, chokepoint routingLowRED
EnergyCold chain, frozen stockpile, controlled-environment local farmsFuel and power prices, grid stabilityLowAMBER
Sea and air freightDetermines whether an alternate source is actually reachable in timeRates, capacity, chokepoint transitsMediumAMBER
Packaging and cold chain consumablesConstrains throughput at processing and distribution nodesResin and paper prices, lead timesHighGREEN
AquafeedDirectly constrains local seafood production, a protein pillar targetFishmeal, alternative protein inputsImprovingGREEN
Aquafeed is included deliberately: SFA and SIT announced a feed and feed ingredient technology centre in June 2026 to drive aquafeed innovation for tropical marine species. Input capability is part of the Grow Local pillar, not separate from it.
OPERATIONS › NODE CAPACITY & INFRASTRUCTURE

Node Capacity & Infrastructure

A diversified source network still funnels through a small number of physical nodes. Ports, cold stores, wholesale markets, inspection and laboratory capacity are the places where a national supply chain can bind even when every source country is healthy — and unlike a source, they cannot be substituted at short notice.

COLD STORE UTILISATION ILLUSTRATIVE
81%
Frozen protein buffer competes for the same space
LAB TURNAROUND HEADROOM
Adequate
Surge capacity is the binding question, not steady state
WHOLESALE THROUGHPUT
Normal
Pasir Panjang and Jurong Fishery Port
SINGLE POINTS OF DEPENDENCE
4
Nodes with no near-term alternative

Node Utilisation Against Capacity

Current utilisation against operating ceiling and against the surge level a severe disruption would demand. Where the surge line exceeds the ceiling, the node is a national constraint regardless of how well diversified the sources are. Prototype figures.

Node Register

NodeFunctionConstraint typeSubstitutable at short noticeWatched signalsState
Sea port food terminalsBulk and containerised arrivalsThroughput, reefer plugsPartiallyBerth waits, reefer availabilityGREEN
Air cargo perishablesChilled substitution modalityBelly and freighter capacityNo, in a peakRates, capacity, festival peaksAMBER
Land checkpointsRegional fresh produce and live animalsClearance throughputNoClearance times, permit stateAMBER
Inspection at entryRisk-based consignment checksOfficer and facility capacityPartiallyInspection rates, detentionsGREEN
Laboratory testingFood safety verificationTurnaround under surgePartially, via recognised labsSample volumes, turnaround, non-compliance rateAMBER
Cold and frozen storageCommercial stock and stockpileSpace and powerBy lease, slowlyUtilisation, temperature, gridAMBER
Wholesale marketFresh produce price formation and distributionThroughput and floor spaceNoArrivals, throughput, pricesGREEN
Fishery portSeafood landing and distributionBerth and handlingNoLandings, vessel callsGREEN
Abattoir and processingLive animal and primary processingLicensed capacityNoThroughput, schedulingAMBER
Local farmsRegenerative productionLand, sea space, productivitySurge over weeksYields, productivity, farm countAMBER
Cold storage carries a compound dependency the platform tracks explicitly: the frozen protein stockpile, commercial frozen stock and any surge substitution to frozen forms all compete for the same space and the same power. A stockpile that cannot stay frozen is not a stockpile.
OPERATIONS › TRADE ROUTES & LOGISTICS

Trade Routes & Logistics

SFA distinguishes route-specific disruptions from those that affect multiple routes and modalities at once. The Suez Canal blockage of 2021 is the first kind; COVID-19 border restrictions across sea and air were the second. A national platform has to score both, because the substitution logic is opposite: a route event is solved by rerouting, a modality event is not.

CHOKEPOINTS WATCHED
6
Sea and air
HIGHEST-RISK CHOKEPOINT
Hormuz
2026 blockade; fertiliser and energy exposure
LONGEST SOURCE LEG
Brazil
Largest meat source, longest sea transit
FALLBACK MODALITY
Air
Fast but capacity-limited at peaks

Transit Time Distribution by Origin

P50 against P90 transit by origin group. The spread is what matters: a long P50 is manageable if it is reliable, but a wide P90 on a source carrying half of an item's volume is a national exposure. Prototype figures.

Chokepoint Watch

ChokepointItems affectedLive indicatorReroute costReroute timeState
Strait of HormuzFertiliser and energy inputs; downstream all cropsTransit counts, war-risk premiaSevere on inputsSeasons, indirectRED
Bab al-Mandab / Red SeaEuropean and Middle Eastern food flowsAIS diversion rate, premia+25–30%+14–20 daysRED
Suez CanalEurope to Asia food exportsTransit counts, backlog, insurance+22%+12–18 daysAMBER
Malacca / Singapore StraitEffectively all sea-borne arrivalsCongestion, berth waitsNot substitutablen/aGREEN
South Atlantic, Brazil legBeef, chicken, porkSailing schedules, port congestion+15%+8 daysAMBER
Land crossings, regionalFresh vegetables, fruit, live animals, eggsClearance times, permit state+5%+1–2 daysGREEN
The Suez precedent gives the calibration: a single grounding blocked the canal for six days, backed up over 400 vessels and disrupted shipping schedules for months. Schedule disruption outlasts the blockage by an order of magnitude, which is why the transit model tracks variance rather than mean.
OPERATIONS › ACCREDITATION & FOOD SAFETY

Accreditation & Food Safety Assurance

There is no food security without food safety. Every substitution the decision engine proposes has to clear accreditation, certification and testing before it is an option at all. The food safety agent therefore holds a veto over the diversification and logistics agents — a faster, cheaper source that is not accredited is not a source.

APPROVED FOR HIGHER-RISK ITEMS SFA 2024
42
Countries and regions: livestock, meat, eggs
NEW SOURCES OPENED, 5 YEARS SFA
10+
For higher food safety concern items
FOODBORNE OUTBREAK RATE SFA 2025
16.8
Per 100,000; down from 22.8 in 2024
ACCREDITATION QUEUE ILLUSTRATIVE
9
Candidates ranked by risk retired

Accreditation Pipeline Ranked by Risk Retired

The platform's most direct operational contribution to the Diversify pillar: ranking candidate source approvals by how much national concentration risk each would retire, rather than by the order applications arrive. Prototype scoring.

Assurance Obligation Register

ObligationApplies toFailure consequenceAutomated checkState
Source accreditationLivestock, meat, egg itemsConsignment ineligible; substitution option voidRegistry status, audit findings, expiryGREEN
Establishment listingIndividual overseas plantsConsignment refused at entryListing status against declared originGREEN
Veterinary health certificationAnimal productsDetention or rejectionCertificate validity, disease-free attestationGREEN
Regionalisation zone evidenceSources with partial outbreaksWhole country lost instead of one zoneZone boundaries against establishment coordinatesAMBER
Import licence and permitAll commercial food importsImport blockedLicence validity, permit declarationsGREEN
Inspection and samplingRisk-based across itemsUnsafe food reaches marketRisk-tiered sampling against arrival profileGREEN
Laboratory testingContaminants, residues, additivesRecall, public health incidentTurnaround, non-compliance clusteringAMBER
Labelling and declarationAll imported foodRecall, enforcement actionDeclaration completenessGREEN
Establishment food safety managementDomestic manufacturing, retail, cateringFoodborne illness outbreakInspection history, demerit accumulationAMBER
The catering channel deserves separate attention. SFA reported that most foodborne illness cases involved catered food and stepped up targeted checks on establishments supplying it. That is a demand-side node where safety risk and supply risk meet, and it belongs in the same register.
OPERATIONS › AGENTIC OPERATIONS

Agentic Operations

Prediction without a prepared response is a newsletter. Each playbook is a bounded, pre-approved workflow that the operating agents assemble on threshold breach and an accountable officer authorises. Every playbook is triggered by a kernel-produced risk value, never by an agent's own judgement, and nothing leaves the agency without a human decision.

PB-01 · ARMED

Source Suspension Response

Trigger: Sanitary or safety event scores HIGH against an approved source.

Agents: Food Safety → Accreditation → Industry Engagement → Communication.

Authority: Officer approval for any external issuance.

PB-02 · ARMED

Export Restriction Response

Trigger: Trade policy agent detects pre-ban language or an announced restriction in a source country.

Agents: Diplomatic → Accreditation → Industry Engagement.

Authority: Ministry escalation for any diplomatic instrument.

PB-03 · ARMED

Regionalisation Activation

Trigger: Outbreak confined to part of a source country where an arrangement exists.

Agents: Food Safety → Diplomatic → Industry Engagement.

Authority: Officer approval; zone evidence attached.

PB-04 · ARMED

Route Disruption Response

Trigger: Chokepoint state turns red or transit P90 breaches the planning window.

Agents: Logistics view → Industry Engagement → Stockpile.

Authority: Officer approval.

PB-05 · STANDBY

Local Surge Activation

Trigger: A perishable, low-substitutability item scores above threshold.

Agents: Local Production → Industry Engagement → Communication.

Authority: Officer approval.

PB-06 · STANDBY

Stockpile Drawdown

Trigger: Coverage clock for a food group falls below the tight band with no substitution path.

Agents: Stockpile → Industry Engagement → Communication.

Authority: Senior management and ministry.

PB-07 · ARMED

Public Confidence Management

Trigger: Demand-side precursors — search interest, purchase acceleration — cross threshold.

Agents: Communication → Industry Engagement.

Authority: Mandatory human review of every message.

PB-08 · ARMED

Enterprise Advisory Issue

Trigger: Any item-level risk crosses threshold with enterprise subscribers exposed.

Agents: Industry Engagement → Food Safety.

Authority: Auto-issue of factual advisories; officer approval for guidance.

PB-09 · STANDBY

Multi-Channel Crisis

Trigger: Two or more food groups below the tight coverage band simultaneously.

Agents: All seven, orchestrated.

Authority: Whole-of-government escalation.

Staging Log

TimestampPlaybookKernel trigger valueStaged actionAuthorityOutcome
T−6h 07:40PB-01HPAI cluster, HIGHCandidate alternate sources ranked; eligibility verifiedOfficerAWAITING DECISION
T−1d 11:15PB-03Zone confined, 0.34Regionalisation evidence pack assembled for source authorityOfficerAUTHORISED
T−2d 16:02PB-08Item risk, MEDIUMFactual advisory issued to subscribed importers and associationsAutoISSUED
T−3d 09:28PB-04Chokepoint REDReroute cost and time modelled; affected consignments identifiedOfficerAUTHORISED
T−5d 13:51PB-07Below triggerMonitoring only; no message draftedAutoHELD
Every log line records the kernel-produced trigger value that justified the staging, and the officer who authorised it. In a government deployment the audit trail is not a compliance overhead — it is the reason a decision can be defended afterwards.
ENTERPRISE › ENTERPRISE RESILIENCE SERVICE

Enterprise Resilience Service

The national platform is not only a command surface for officers. Every risk value the kernel produces is also a decision input for a Singapore enterprise — the importer deciding whether to hold cover, the manufacturer deciding whether to reformulate, the caterer deciding whether to reprice a contract, the supermarket deciding what to put on the shelf. The state sees the whole picture; enterprises act inside it. Connecting the two is where national resilience is actually built.

ADDRESSABLE ENTERPRISES ILLUSTRATIVE
~9,000
Importers, traders, manufacturers, wholesalers, retailers, caterers
WARNING PASSED DOWNSTREAM
6.4wk
Mean lead time available to an enterprise
MARGINAL COST TO SERVE
Near zero
The risk value is already computed for the state
NATIONAL EFFECT
Compounding
Thousands of prepared firms is national resilience

Who the Service Serves

Importers & traders

Hold the primary substitution decision. Need to know which alternate sources are accredited, available and arrivable — before the primary source closes, not after.

Source risk alertsSubstitution setAccreditation pipeline

Food manufacturers

Hold reformulation and ingredient-substitution options that take weeks to qualify. Lead time is worth more to them than to anyone else in the chain.

Ingredient riskInput cost outlookQualification lead

Cold store & logistics operators

Need to anticipate space, power and modality demand before a surge, not during one. A frozen substitution wave is a warehouse problem before it is a shelf problem.

Surge forecastRoute disruptionModality shift

Wholesalers & distributors

Sit where arrivals meet demand. Early warning converts a scramble into a scheduled reallocation across customers.

Arrival outlookPrice formationAllocation planning

Retailers & supermarkets

Face the demand-side shock directly. Panic-buying precursors and form-substitution guidance are worth more to them than any supply metric.

Demand surge signalsForm substitutionShelf planning

F&B, caterers & hawkers

The least resourced and the most exposed. In 2022 hawkers and stallholders switched sources with association support — the platform makes that switch informed and early rather than reactive.

Item alertsMenu substitutionAssociation channel

Service Tiers

Reach against depth across three service tiers. The public-good tier reaches everyone at almost no marginal cost; the enterprise tier is where firms with real substitution decisions get item-specific, firm-specific analysis. Prototype sizing.
TIER 1 · PUBLIC GOOD

National Food Risk Advisory

Free item-level risk state and early warning, published to all licensed food businesses. The same information SFA already issues as trade circulars, but continuous, scored, and item-specific rather than event-driven.

DELIVERS
Item risk statesSource alertsAdvisory feed
NO CHARGE
TIER 2 · SECTOR

Association & Sector Dashboards

Sector-level views for trade associations and chambers — meat traders, importers, food manufacturers, caterers — with substitution readiness and collective sourcing intelligence. Associations were the delivery channel in 2022 and 2023; this gives them the instrument.

DELIVERS
Sector dashboardSubstitution setSourcing trip targeting
COST RECOVERY
TIER 3 · ENTERPRISE

Firm-Level Resilience Desk

API and dashboard scoped to a firm's own declared import portfolio: exposure by item and source, business continuity score, cost-of-switch estimates, and alerting tuned to what that firm actually buys.

DELIVERS
Portfolio exposureBCP scoreCost of switchAPI
SUBSCRIPTION

Why the State Should Provide It

ArgumentReasoning
The information is already producedThe kernel computes item-level risk for national purposes regardless. Distributing it to the firms who act on it is nearly free at the margin and multiplies its effect.
Only the state sees the whole boardNo individual importer can see 180-plus source countries, the accreditation registry, disease notifications and chokepoint state at once. This is a natural public-information function.
Business continuity plans need inputsSFA already works with industry to strengthen supply networks through business continuity plans. A BCP built on a firm's own guesswork is weaker than one built on the national risk picture.
Diversification is executed by firmsThe state accredits sources; enterprises actually buy from them. An accreditation nobody uses retires no risk. Telling firms which new sources matter for their portfolio converts approvals into flows.
It reduces the demand-side shockRetailers and F&B operators who know a substitution is coming can stage it calmly. Panic buying is often a response to visible, unexplained absence rather than to actual scarcity.
It creates a feedback loopFirms report what they actually sourced and what it cost. That closes the loop between the model's predicted substitution set and the real one, and improves the national picture.

Enterprise Value by Segment

SegmentDecision the service improvesCurrent practiceWith the platformValue
ImporterWhen to open a second source and how much cover to holdReactive; triggered by a supplier's own noticeScored source risk weeks ahead; alternate set pre-verified as accreditedHIGH
ManufacturerWhen to qualify an alternate ingredient originQualification begins after disruption; takes weeksQualification begins during the warning windowHIGH
Cold store operatorSpace and power planning for a frozen substitution waveDemand appears without noticeSurge forecast by food group with lead timeMEDIUM
WholesalerAllocation across customers during a shortfallFirst-come allocation under pressurePlanned allocation against a forecast arrival profileMEDIUM
RetailerWhat to stock, what to substitute, what to communicateShelf gap first, explanation secondSubstitution staged and explained before the gap appearsHIGH
Caterer / F&BMenu and contract pricing under input volatilityMargin absorbed or contract renegotiated lateInput outlook by item with enough lead to reprice properlyHIGH
Hawker / small operatorWhether and when to switch supplier or formWord of mouth and supplier noticePlain-language item alerts through association channelsMEDIUM
ENTERPRISE › BCP & SUBSTITUTION DESK

Business Continuity & Substitution Desk

The operational instrument an enterprise actually uses. A firm declares its import portfolio; the desk returns its exposure, its business continuity score, and a ranked, pre-verified substitution set for every item it buys. When a risk value crosses threshold, the firm receives the alert with its own options already attached.

SAMPLE FIRM · BCP SCORE ILLUSTRATIVE
62/100
Mid-size chilled and frozen meat importer
SINGLE-SOURCE ITEMS
3
Of 11 items in the declared portfolio
VERIFIED ALTERNATES
17
Accredited and currently exporting
WARM LINKS HELD
5/17
Relationships that actually exist today

Business Continuity Score — Components

A firm's continuity position across six dimensions against the sector benchmark. The gap on warm supplier links is the most common weakness the desk surfaces: firms know their alternates exist on paper but have never transacted with them. Prototype scoring for a representative firm.

Substitution Desk — Worked Output

What a subscribed importer sees when the kernel scores a poultry source event as HIGH. Every alternate is checked for eligibility before it is shown — an unaccredited option is not an option.

RankAlternate sourceAccreditedCurrently exportingEst. lead timeCost deltaFirm has warm linkRecommendation
1Regional chilled, existing approved sourceYESYes2 weeks+8%YESActivate first; lowest friction
2Frozen, long-haul approved sourceYESYes5 weeks−4%YESCheapest; form change required at the customer end
3Newly approved European sourceYESLimited volume7 weeks+19%NOOpen the relationship now, before it is needed
4Newly approved South American sourceYESEstablishing9 weeks+14%NOCandidate for the next sourcing mission
5Non-accredited alternativeNOn/an/an/aNONot an option; flagged to SFA as an accreditation candidate
Row five is the loop closing in the other direction. When enterprises repeatedly want a source that is not accredited, that demand signal belongs in the national accreditation pipeline — which is exactly how SFA already works, connecting local importers with overseas businesses and facilitating engagements such as the 2026 Meat Traders' Association visit to Peru, Paraguay and Chile.

From Advice to Instrument

SFA guidance to businessWhat it requires to be actionableDesk function that supplies it
Strengthen business continuity plansA measured starting position and a defined gapBCP score across six dimensions, benchmarked to sector
Keep options open with multiple suppliersKnowing which alternates are eligible and live, not just listedVerified substitution set, refreshed continuously
Maintain warm linksKnowing which relationships are cold and which matter mostWarm-link gap analysis ranked by exposure retired
Be prepared, not reactiveWarning early enough for qualification and contractingItem-level alerts with weeks rather than days of lead
Stay flexible on form and sourceUnderstanding the cost and lead time of each form changeCost-of-switch estimates by form and origin
Support local produceKnowing when local supply is the fastest available bufferLocal surge availability shown alongside import alternates
Nothing in the left column is new. The entire contribution of this view is to make each line measurable for an individual firm — which is what turns published national guidance into a change in behaviour.

National Effect of Enterprise Adoption

Adoption levelEnterprise behaviourNational consequenceMeasurable as
Low — advisory onlyFirms react when their own supplier notifies themSubstitution begins after the shortfall is visibleShelf gaps, price spikes, public concern
Moderate — sector dashboardsAssociations coordinate switching across membersSubstitution begins in parallel across a sectorShorter disruption duration
High — firm-level desksFirms hold pre-verified alternates and warm linksSubstitution begins during the warning window, before the event bindsDisruptions that never become visible to consumers
Full — with feedback loopFirms report actual sourcing and cost outcomesThe national model learns which substitutions are realAccreditation pipeline ranked by genuine demand
The strategic point for a national platform: resilience is not held by the state alone. A thousand prepared firms absorb a shock that no stockpile can cover, and they do it without drawing on any public buffer at all.
PLATFORM › ARCHITECTURE STACK

Platform Architecture

Eight layers. The granted patent occupies Layer 4 and is the immutable core. The analytical moat sits in Layers 3 and 5, where the national ontology and the custom models hold relationships no general-purpose system can assemble. The enterprise service is a distribution surface on Layer 8, not a separate product.

L8
National C3 Surface & Enterprise Distribution
Officer command surface, decision record, escalation and audit trail — plus the advisory, sector and enterprise tiers that push the same scored risk out to Singapore firms.
NEW IP
L7
Agentic Staging Layer
Accreditation, industry engagement, diplomatic, stockpile, local production, food safety and communication agents assembling executable responses for officer authorisation.
FILING CANDIDATE
L6
Inference & Decision Engine
Propagates a scored event across the national ontology to every affected item, source, importer, route and stockpile line, then generates and prices the pillar option set.
FILING CANDIDATE
L5
Custom AI & Data Model Layer
Six models: disease and trade restriction propagation, corridor transit, substitution feasibility, consumption and confidence, climate and local production, input and price.
MOAT
L4
Patented Risk Determination Kernel
Risk value computation with positive and negative scores, impact percentile against threshold, high / intermediate / low classification, supply-side versus demand-side attribution. The single authoritative source of a risk value in the platform.
SG10202109991Y
L3
National Food Ontology & Knowledge Graph
Food item, source country, establishment, importer, consignment, route, chokepoint, disease, jurisdiction, stockpile line, local farm and consumption group. The semantic spine that lets one Malaysian rainfall event score against eggs, vegetables, fruit and seafood at once.
MOAT
L2
Data Layer & Sensing Substrate
Import declarations and volumes, accreditation registry, inspection and laboratory results, farm production returns, stockpile records, cold store and wholesale telemetry, plus external veterinary, climate, maritime and commodity feeds.
CLAIM ELEMENT
L1
LLM Layer
Multilingual extraction and reasoning across the languages of Singapore's source countries. Infers a disease or policy event from description before formal labelling — the capability a keyword classifier structurally cannot provide.
FOUNDATION
Continuity note: this is the same eight-layer stack used in the AR livestock deployment, with the ontology and data layers re-instantiated for a national import system. The kernel, its position and its governing rule are unchanged.

Capability Delta Against the 2021 Filing

CapabilityAs disclosed, 2020–21Deployed architecture, 2026New subject matter
Event extractionKeyword dictionary — pandemic, natural disaster, geopolitical tension, corporate bankruptcyMultilingual LLM inference from unlabelled veterinary, policy and market descriptionPARTIAL
Credibility filteringBehavioural-versus-text score comparison using gradient boosting and neural classifiersAdversarial multi-agent assert / falsify / reliability architectureYES
Risk aggregationIndependent event scoring, summedCorrelated country exposure across multiple food items scored as one eventYES
Threshold logicPreset or user-set threshold valueThreshold modulated by remaining stockpile coverage days for the affected food groupYES
Implication reasoningNot disclosedInference engine propagating a scored event across a national ontology to derive pillar sequencingYES
Response activationResponse plan activated by machine learning algorithmsMulti-agent staging with mandatory officer authorisation and a food safety vetoPARTIAL
Growth monitoringPlant growth monitoring at physical farmsLocal farm production and productivity monitoring as the regenerative offset in the risk valueRE-INSTANTIATION
Command-and-control centreExpressly disclosed in the specificationRealised as a national command surface with enterprise distributionAS DISCLOSED
PLATFORM › NATIONAL FOOD DATA SPINE

National Food Data Spine

The platform's durable asset is not the model. It is the linked record that connects every consignment to its source establishment, its item, its importer, its route, its inspection result and its contribution to national supply. Public data explains the world; the spine explains Singapore's position in it, and nobody outside government can assemble it.

Consignment Identity Chain

Source Country
APPROVAL STATUS
Establishment
ACCREDITATION ID
Food Item
CLASSIFICATION
Certification
HEALTH CERT · ZONE
Route
MODALITY · CHOKEPOINTS
Importer
LICENCE · PORTFOLIO
Declaration
PERMIT · VOLUME
Inspection
OUTCOME · SAMPLING
Lab Result
TEST · TURNAROUND
Storage
COLD STORE · STOCKPILE
Distribution
WHOLESALE · TRADE
Consumption
FOOD GROUP CONTRIBUTION

Analytical Capability Compounding

Linked consignment-years accumulated against substitution model accuracy. The substitution feasibility model cannot outperform an expert's judgement until several years of linked history exist — which is why the spine is a phase two priority even though its payoff arrives later.

What the spine feeds

Substitution feasibility scoring. Concentration and correlation analysis by country and item. Accreditation pipeline ranking. Enterprise portfolio exposure. Regionalisation zone evidence. Inspection targeting.

Why it is defensible

Trade statistics show what arrived. The spine shows which establishment sent it, under which certificate, through which chokepoint, to which importer, and whether the substitute the model proposed was actually usable. No commercial provider can replicate it.

What it must not become

A commercial intelligence product about individual firms. Enterprise-facing outputs must be scoped to a firm's own portfolio and to aggregate national state — never to a competitor's positions. Get this wrong and the enterprise tier destroys the trust the whole platform depends on.

Data Governance Requirements

RequirementWhy it is non-negotiableDesign consequence
Firm-level confidentialityImport declarations reveal commercial positionsEnterprise tier returns only the firm's own data plus national aggregates
Aggregation thresholdsA small-n aggregate can identify a single firmMinimum counts enforced before any aggregate is published
Provenance on every risk valueA national decision has to be explainable afterwardsEvery score carries its evidence chain and model version
Source data stays authoritativeThe platform must not become a competing record of truthReads from systems of record; never overwrites them
External data separationOpen-source intelligence has different reliability to administrative dataVerification layer scores and labels every input by class
Cross-agency access controlWhole-of-government value requires sharing without leakageRole-scoped views; national state shared, firm detail not
PLATFORM › PATENT & IP POSITION

Patent & IP Position

The platform rests on a granted Singapore patent whose specification describes, in terms, a command-and-control centre for monitoring and controlling food supply and demand — including the worked example of a Singaporean import-substitution decision across rice and vegetables. A national food security deployment is not an adaptation of this invention. It is the use case the specification anticipated.

GRANTED PATENT
SG 10202109991Y
Method and system for risk determination in a food supply chain
PRIORITY / GRANT
2020 → 2023
Priority 11 Sep 2020 · filed 10 Sep 2021 · granted 6 Dec 2023
ORIGINAL TITLE
Food Security Exchange
Priority application, filed with IPOS on 11 Sep 2020

The Specification Already Describes This Deployment

Command-and-control centreFigure 3 of the specification describes an establishment performing monitoring and controlling operations related to demand and supply in the food supply chain, issuing warnings to a distribution network and activating response plans across a supplier network.
National import substitutionThe specification's worked example is Singapore importing vegetables from alternative countries including the United States, India, China, Vietnam, Australia and Thailand, and rice from alternative origins — classified as intermediate and low risk precisely because substitution exists.
Global food supply chainThe specification states expressly that risk determination may be performed for a global food supply chain, with a supplier network associated with multiple countries and import and export operations monitored accordingly.
Regional trigger eventsThe specification's illustrative trigger is a disaster affecting Mekong delta rice fields, classified as high risk with severe effect on rice supply in surrounding regions — the same regional concentration risk SFA identifies today.
This is an unusually strong alignment between a granted claim set and a national mandate. The priority application was titled Food Security Exchange System and filed with IPOS in September 2020.

Claim Element Mapping to the National Deployment

Claimed elementNational instantiationCoverage
Trigger event data by machine learning for NLPDisease notifications, export restrictions, climate events, chokepoint incidents, extracted multilinguallyDIRECT
Keywords: pandemic, natural disaster, geopolitical tension, corporate bankruptcyThe same four classes map onto SFA's published risk taxonomy almost exactlyDIRECT
Geospatial correlation in real timeOutbreak and event zones intersected against accredited establishment and source region coordinatesDIRECT
Weather forecast dataMekong hydrology, ENSO state, regional rainfall against source-region exposureDIRECT
Natural disaster prediction moduleFlood, drought and heat event identification against producing regionsDIRECT
Growth monitoring at physical farmsLocal farm production and productivity as the regenerative bufferRE-INSTANTIATED
Risk value with positive and negative scoresNegative: volume at risk. Positive: regionalisation, alternate capacity, local surge, stockpileDIRECT
Impact percentile against thresholdThreshold modulated by remaining coverage days for the food groupDIRECT
High / intermediate / low classificationPer item, per source, per corridor — the specification's own rice and vegetable exampleDIRECT
Supply-side / demand-side classificationSource disruption versus panic buying and consumption surgeDIRECT
Response plan: change suppliers, adjust ordersDiversify, partnerships, stockpile, grow local — sequenced and pricedDIRECT
Alerting via IoT-based sensorsCold store, wholesale and stockpile telemetry; local farm sensingREQUIRES DEPLOYMENT
The characterising portion of both independent claims requires alerting via IoT-based sensors. In a national deployment the natural instantiation is cold chain, storage and local farm telemetry — which the platform needs for its own purposes regardless.

Fresh Filing Candidates

CANDIDATE 01 · STRONGEST

Buffer-Modulated Risk Threshold

Risk impact threshold dynamically set by the remaining stockpile coverage of the specific resource at risk, so identical raw risk escalates earlier as the buffer thins. Absent from the 2020 disclosure.

NOVEL
CANDIDATE 02 · STRONG

Correlated Multi-Item Source Scoring

Computing a single risk value across multiple distinct supply chains sharing a common source geography, rather than scoring each chain independently.

NOVEL
CANDIDATE 03 · STRONG

Adversarial Verification Architecture

Multi-agent assert, falsify and source-reliability scoring for supply chain event verification. Extends the disclosed credibility sub-module without being disclosed by it.

NOVEL
CANDIDATE 04 · CONDITIONAL

Eligibility-Gated Substitution Ranking

Ranking substitution options by simultaneous regulatory eligibility, availability, cost and arrival feasibility, with an assurance veto. Closer to the disclosed response plan step.

CONTESTED

Public Sector Licensing Structure

Licence, not assignmentA field-of-use licence to the agency for national food security purposes, with the patent retained by the proprietor. The agency gets certainty of use; the asset remains licensable into other jurisdictions.
Government rightsSingapore's patent legislation provides for government use of patented inventions in specified circumstances. A negotiated licence on clear terms is preferable for both parties to relying on those provisions.
ImprovementsInventions arising from the deployment filed by the proprietor and licensed back on the same terms, with the agency holding rights of use in perpetuity for the national mandate.
Data ownershipAll national data, the ontology content built from it, and every model trained on it remain the property of the agency. The licence covers the method, never the data.
Enterprise tier revenueIf the enterprise tier charges, the agency sets pricing and retains receipts. The proprietor's return is the licence, not a share of a public service.
ExitPerpetual paid-up rights on termination for anything already deployed, so no national capability is ever contingent on a commercial relationship continuing.
VALUE › PUBLIC VALUE & DESTROYERS

Public Value Drivers & Destroyers

Where the platform creates public value, and where it destroys it if built wrongly. Both sides are stated, because in government deployments the destroyers are the more common outcome and their costs are borne publicly.

Public Value Drivers

D1 · Detection compression
The same correct response, started weeks earlier. The single largest source of value in the system.
HIGH
D2 · Correlated exposure visibility
One country event scored across every item it touches, instead of item-by-item reports that each look manageable.
HIGH
D3 · Accreditation prioritisation
Source approvals ranked by concentration risk retired, converting effort into resilience rather than throughput.
HIGH
D4 · Enterprise multiplication
Thousands of prepared firms absorbing shocks without drawing on any public buffer.
HIGH
D5 · Stockpile efficiency
Buffer sized and rotated against modelled risk rather than a fixed rule, at real carrying cost saved.
MED
D6 · Demand-side management
Panic buying anticipated and defused rather than observed. Cheap intervention, high public value.
HIGH
D7 · Decision auditability
Every national decision carries its evidence chain, which matters when it is examined afterwards.
MED
D8 · Diplomatic timing
Knowing when to activate an agreement is worth more than having one. The 2023 rice exemption is the model.
HIGH
D9 · Exportable capability
A working reference implementation other import-dependent states will want, built on Singapore-held IP.
MED

Public Value Destroyers

X1 · False positives in public
A wrong national warning costs public trust and can itself cause the panic buying it predicted. Verification is not optional.
SEVERE
X2 · Alerts without authority
A system that warns but cannot route into a decision becomes a report nobody reads.
SEVERE
X3 · Kernel bypass
If an LLM or agent originates a risk verdict, the national number has no auditable basis and the patent stops being the asset.
SEVERE
X4 · Commercial confidentiality breach
Enterprise tier leakage of one firm's positions to another ends industry cooperation permanently.
SEVERE
X5 · Automation of a public message
Any communication issued without human authorisation is a reputational incident waiting to happen.
SEVERE
X6 · Dashboard without ontology
Views over disconnected data cannot see correlated exposure. Layer 3 is the whole point.
HIGH
X7 · Duplicating systems of record
A second competing version of import or accreditation data is worse than none.
HIGH
X8 · Model drift after regime change
Source behaviour after a ban is a different process than before it. Unmonitored models mislead confidently.
HIGH
X9 · Vendor dependency
A national capability contingent on a commercial relationship continuing is not a national capability.
HIGH

Value Contribution by Driver

Illustrative relative contribution to modelled public value, prototype figures only.
VALUE › DEPLOYMENT FEASIBILITY

Deployment Feasibility

Four phases, each independently useful so the programme can stop at any boundary without stranding value. Within them runs the same seven-step build sequence, which always begins with the patented core and ends with the command surface and its enterprise distribution.

PHASE 1 · 0–5 MONTHS

Sense & Score

Signal scanner, verification layer, import data explorer, source and corridor board, precedent-calibrated thresholds. Officer pilot on two food groups.

Build steps 1–2Patent worked
PHASE 2 · 5–12 MONTHS

Model & Predict

Six models trained on the national data spine. Coverage clock, disease and regionalisation, and consumption intelligence operational across all food groups.

Build steps 3–4Moat begins
PHASE 3 · 12–20 MONTHS

Reason & Stage

Inference and decision engines live. Agentic staging with officer authorisation and food safety veto. Four pillars command in production.

Build steps 5–6Loop closes
PHASE 4 · 20–30 MONTHS

Distribute

Enterprise resilience service across the three tiers, whole-of-government interfaces, and the international reference implementation.

Build step 7Enterprise multiplication

Seven-Step Build Sequence

StepBuildWhat it deliversPhaseDepends on
01Patent coreThe risk determination kernel as the platform's only risk authorityPhase 1
02National data connectionImport declarations, accreditation registry, inspection and lab results, farm returns, stockpile recordsPhase 1Step 1
03AI modelsDisease, transit, substitution, consumption, climate and input forecasting feeding the kernelPhase 2Step 2
04LLM layerMultilingual external event intelligence and the officer conversational interfacePhase 2Step 2
05Inference engineRisk plus models plus national ontology resolved into consequencePhase 3Steps 3–4
06Agentic stagingAccreditation, engagement, diplomatic, stockpile and communication workflows prepared for authorisationPhase 3Step 5
07C3 surface & enterprise tiersOne national command surface, and the same intelligence distributed to Singapore firmsPhase 4Steps 1–6
Step 1 comes first without exception. Building models or agents before the kernel produces a platform in which the patent is retrofitted rather than foundational — the failure mode X3 describes, and the one that leaves a national risk number with no defensible basis.

Capability and Investment Trajectory

Cumulative capability against cumulative investment, illustrative. The inflection at Phase 2 is where the national data spine begins to compound; the inflection at Phase 4 is enterprise multiplication.

Feasibility Assessment

DimensionAssessmentBasisRating
Technical feasibilityEvery component is proven capability. The work is orchestration and domain modelling, not research.Existing LLM, agent and geospatial stacksHIGH
Data availabilityExternal sources are public. The proprietary layer is the agency's own administrative record, which already exists.Signal scanner and data spine registersHIGH
Institutional adoptionRequires officers to act on model output under time pressure, and a clear authority design. This is the binding constraint, not technology.Playbook authorisation modelMEDIUM
Legal and governanceData governance, confidentiality and auditability requirements are demanding but well understood.Data spine governance registerMEDIUM
Industry participationEnterprise tiers require industry trust in confidentiality. Association channels already exist and have been used in past responses.2022 and 2023 industry responsesMEDIUM
IP positionGranted patent covers the scoring core and expressly contemplates a food supply command-and-control centre. Three strong candidates for fresh filing.Patent & IP Position viewHIGH
VALUE › WHOLE-OF-GOVERNMENT & ASEAN

Whole-of-Government & International Reference

Food security is not held by one agency. Trade policy, foreign affairs, transport, enterprise development, health and national security all hold instruments that the same scored risk should route into. And a working national platform becomes a reference implementation other import-dependent states will want — built on Singapore-held intellectual property.

Whole-of-Government Interfaces

PartnerInstrument they holdWhat the platform gives themTrigger
Trade & industryTrade agreements, market access, price and competition policyEarly notice of source-country restrictions and their national volume effectTrade policy risk above threshold
Foreign affairsDiplomatic engagement, bilateral agreementsWhen to activate an agreement, and what it is worth in tonnes and daysPolitical rather than physical constraint detected
Transport & maritimePort capacity, shipping and air cargo policyChokepoint state and modality substitution demandRoute risk above threshold
Enterprise developmentBusiness support, internationalisation, sourcing missionsWhich source countries and firms to target on the next missionAccreditation pipeline ranking
HealthFood safety, foodborne illness surveillance, nutritionEmerging contaminant and outbreak signals from source countriesSafety risk above threshold
National security & crisis managementWhole-of-government crisis coordinationCoverage clock and multi-channel crisis stateTwo or more groups below the tight band
Environment & climateClimate adaptation, resource planningRegional climate transmission into national food exposureClimate risk above threshold
Statistics & dataNational data infrastructure and standardsA linked food supply data spine built to whole-of-government standardsContinuous

ASEAN and International Reference

Transferability scored on structural similarity to Singapore's position, evidence transferability from the national reference implementation, and likely institutional appetite. Prototype scoring.

ASEAN food security cooperation

An ASEAN ministers' statement on safeguarding food security and supply chains already frames the regional commitment. A working risk platform gives that commitment an operational layer — shared early warning on transboundary disease, climate events and export restrictions.

NEAREST FIT

Import-dependent city-states and small states

Structurally identical problem: high import dependence, limited land, concentrated source exposure, and a stockpile as the buffer of last resort. The platform transfers with the ontology re-instantiated and little else changed.

HIGH TRANSFER

Gulf states

Comparable import dependence and stockpiling posture, with additional route concentration through the very chokepoints the platform already watches. Strong appetite for food security infrastructure.

HIGH TRANSFER

Bilateral partners

Agreements such as the essential supplies agreement with New Zealand and the rice trade agreements with Cambodia, Thailand and Vietnam work better when both sides can see the same risk picture and the same activation triggers.

ADJACENT

Multilateral bodies

Regional and international food security mechanisms need a common analytical basis for early warning. A method with a granted patent and a working national reference is a credible contribution.

LONG CYCLE

Commercial adjacencies

Insurance and trade finance both need scored, auditable supply chain risk. These are licensing routes for the underlying method that do not touch national data at all.

ADJACENT

What a Working National Reference Establishes

Proof pointDemonstrated byTransfers to
The method works at national scale180-plus source countries, eight food groups, one scored pictureEvery import-dependent state
Correlated exposure is real and scoreableOne country event scored across four items simultaneouslyAny multi-item sourcing portfolio
Lead time is achievable, not theoreticalHistorical replay against Singapore's own documented disruptionsEvery subsequent conversation
The state–enterprise loop closesFirms acting on national risk values and reporting outcomes backAny economy where diversification is executed by private firms
Governance can be satisfiedOfficer authority, audit trail, confidentiality and verification all designed inEvery public sector buyer
The patent is workedThe command-and-control centre the specification describes, built and operatingLicensing and enforcement posture generally
FS RISK › OVERVIEW

Food Security Risk — National Overview

Food security is a national security concern. Food supply disruptions pose strategic risks, and governments and industries need to understand their food supply chain exposures and vulnerabilities before an event, not after it. This section quantifies where Singapore's exposure actually sits — and the answer is not evenly distributed.

SINGLE-COUNTRY EXPOSURE
Malaysia
Major source across six or more food items
LEAFY VEGETABLES FROM MALAYSIA SFA
69%
Of imported leafy vegetables
EGGS FROM MALAYSIA SFA / CNA
52%
Down from 72% in 2019
LIVE PIGS FROM MALAYSIA
100%
One state, one farm, one importer
SINGLE POINTS OF FAILURE
7
Identified across the national chain

FS Risk by Domain

Exposure, likelihood and detectability across seven risk domains. The pattern that matters for resource allocation: the domains with the highest exposure are not the ones with the worst detectability, so the binding constraint is analytical capacity rather than data availability. Prototype scoring.

The Three Risk Classes and How They Bind

CLASS 01 · MACRO-ENVIRONMENTAL

Climate, Pandemics, Natural Disasters

Affects an entire region at once. Rice is mainly produced in the Mekong region, and harvests take months or years to recover once disrupted. Prolonged hot and dry conditions during El Niño span multiple growing cycles.

Substitution fails when the whole producing region is affected. This is the class where stockpile leads.
CLASS 02 · SOCIO-POLITICAL

Policy Decisions, Export Bans, Transport Disruption

Political turmoil, decisions to ban exports, and transport disruption. Singapore has been directly hit twice in recent years — the Malaysian live chicken export ban in 2022 and India's non-basmati rice ban in 2023.

The constraint is political, so the diplomatic instrument outranks commercial substitution. Singapore obtained a rice exemption in 2023.
CLASS 03 · CORPORATE & BIOSECURITY

Supplier Failure, Disease, Scandals

A supplier in corporate trouble, a scandal, or a biosecurity breach. Unlike crops, livestock take longer to breed and grow — a single disease incursion may disrupt meat supplies over many months or years.

This is the class where a single establishment can remove a national supply line overnight, as ASF at Pulau Bulan did in 2023.

Five Years of Repeated Testing

Singapore's own ministerial account is that over the past five years the food supply has been repeatedly tested by events beyond its control, and that each time the country tided through with the collective support of importers, retailers and consumers who switched to alternative options.

YearEventItem affectedMechanismClass
2020COVID-19 border measuresMultiple, notably eggsLockdowns and export restrictions across sea and airMACRO
2021Suez Canal blockageEuropean-origin foodSix-day closure, 400-plus vessel backlog, months of schedule disruptionSOCIO-POLITICAL
2022Malaysia live chicken export banFresh chickenDomestic shortage driven by feed cost escalationSOCIO-POLITICAL
2022Heavy rainfall, Melaka and Cameron HighlandsLeafy vegetablesYield loss plus quality degradation and faster spoilageMACRO
2022Russia–Ukraine conflictEggs, via feedGrain disruption raised chicken feed costs, then egg pricesSOCIO-POLITICAL
2023ASF at Pulau Bulan, IndonesiaFresh porkLive pig imports halted; a major source removed overnightBIOSECURITY
2023India non-basmati rice export banRiceWorld's largest rice exporter restricts exportsSOCIO-POLITICAL
2025HPAI in Brazil; ASF in MalaysiaPoultry; porkRegionalisation preserved Brazilian supply; Malaysian culling raised live pig pricesBIOSECURITY
2025–26Rolling HPAI suspensionsPoultry and poultry productsNear-monthly suspension and lifting cycles across a dozen source countriesBIOSECURITY
2026Strait of Hormuz blockadeAll crops, via fertiliserEnergy and fertiliser flows disrupted; yields affected over seasonsSOCIO-POLITICAL
Ten material events in six years. That is not a tail risk profile — it is a steady-state operating condition, and it is the reason continuous scored monitoring is a different proposition from periodic review.
FS RISK › MALAYSIA DEPENDENCY

The Malaysia Dependency Risk

The single largest concentration in Singapore's food system. Malaysia is a major source across at least six food items simultaneously, shares a land border with two fixed crossings, and is itself exposed to the same climate, disease and fertiliser shocks that would trigger a Singapore shortage. A per-item view of the import portfolio does not show this. A country-level view does.

Diversification counted at the country level looks strong: over 180 source countries and regions. Diversification weighted by volume tells a different story. One of Singapore's largest sources of food is Malaysia, which supplies more than half its eggs, about 40 per cent of its fruits and vegetables, and a sizeable amount of meat and seafood — and, per SFA data, 69 per cent of imported leafy vegetables.

LEAFY VEGETABLES SFA
69%
Cameron Highlands concentrated
SHELL EGGS CNA / SFA
52%
Was 72% in 2019
ALL VEGETABLES CNA / SFA
42%
Of 567,900 t, 2021
FRUIT CNA / SFA
40%
Of 433,100 t, 2021
CHICKEN SFA
34%
Fresh chicken almost entirely Malaysian

Malaysia's Share Across the Import Portfolio

Malaysia's share by food item, against the item's total import volume. The bubble is not one exposure — it is six or more exposures that move together, because they share a single geography, a single climate, a single regulatory authority and a single land corridor.

Why the Exposure Is Correlated, Not Diversified

Shared factorWhy it correlates the itemsDocumented instanceItems bound together
Single climate zoneOne monsoon or rainfall event covers the growing regions for several items at onceProlonged heavy rain in Melaka and Cameron Highlands, 2022, depleted yields and degraded qualityVegetables, fruit
Single regulatory authorityOne export policy decision applies across items simultaneouslyLive chicken export ban, June 2022, to address domestic shortageChicken, potentially eggs
Shared disease geographyTransboundary animal disease spreads across the peninsula and affects several protein linesASF across Malaysia; pork self-sufficiency fell from 93.4% in 2021 to 69.6% in 2023Pork, live pigs
Shared upstream inputsMalaysia imports over 60% of its mineral fertiliser; an input shock hits all its crops together2026 Hormuz disruption and fertiliser export curbs raising Malaysian production costsVegetables, fruit
Single land corridorTwo fixed crossings carry the fresh, perishable, time-critical flowsBorder friction and clearance delays affect all land-borne fresh produce at onceVegetables, fruit, eggs, live animals
Shared feed dependencyMalaysian and Singaporean egg and poultry producers buy the same imported feedRussia–Ukraine grain disruption raised feed costs and egg prices in both countriesEggs, poultry — including local production
The last row is the most uncomfortable. Local egg production is Singapore's buffer against Malaysian egg supply — but both depend on the same imported feed. A feed shock degrades the exposure and the hedge simultaneously.

Cascade Model — One Upstream Shock, Multiple Singapore Items

The 2026 fertiliser sequence traced through: Hormuz disruption and export curbs raise Malaysian fertiliser costs, which raise production costs for fertigation-dependent crops, which reduce planted area and yield, which reduces Singapore-bound volume and quality. Prototype modelling of a documented, in-progress chain.

The Live Pig Chain — Concentration in Four Successive Layers

The clearest illustration of why country-level diversification statistics understate real risk. Every layer of this chain has a single point.

LayerCountDetailIf it fails
Source country for live pigs1Malaysia only, following the halt of Pulau Bulan supply in April 2023No live pigs at all
Source state1Sarawak is the only Malaysian state licensed to export live pigs to SingaporeNo live pigs at all
Source farm1One anchor farm shipped over 121,000 live pigs in 2024 — about 8.2% of Singapore's total pork importsFresh pork line ends
Singapore importer1A single importer brings in roughly 3,000 live pigs a week, twice weekly under veterinary controlFresh pork line ends
Abattoir1One licensed abattoir in Jurong, with same-day butchering and distributionFresh pork line ends
Chilled and frozen fallback20+Chilled from Australia and Canada; frozen from Brazil, the US, Spain and ChinaVolume is safe; form changes
The 2023 response worked precisely because the last row existed. The lesson the platform encodes: a supply line can be simultaneously volume-secure and form-fragile. Scoring only tonnage would have shown this chain as low risk right up to the day it stopped.

Historical Precedent for a Total Malaysian Halt

1999
The live pig trade between Malaysia and Singapore is halted following the Nipah virus outbreak.
1999–2017
Eighteen years without live pig imports from Malaysia. Singapore's supply shifted entirely to Pulau Bulan, Indonesia, which had begun exports in 1987.
NOV 2017
Live pig imports resume — exclusively from Sarawak.
APR 2023
ASF detected in a consignment from Pulau Bulan. Singapore halts that source. Sarawak becomes the sole live pig origin.
2026
SFA and the Animal and Veterinary Service work with Indonesian authorities on the possible resumption of Pulau Bulan imports — restoring a second origin.
An eighteen-year outage is the base rate for a biosecurity-driven livestock trade halt between these two countries. Any risk model that treats a Malaysian supply line as recoverable in months is calibrated against the wrong history.

What the Platform Does About It

CapabilityApplied to the Malaysia exposurePillar it serves
Country-level correlated scoringOne Malaysian trigger event scores against leafy vegetables, all vegetables, fruit, eggs, chicken, seafood and live pigs in a single risk value, rather than seven separate low readingsDiversify
Upstream fertiliser monitoringWatches Malaysia's fertiliser import position as a leading indicator of Singapore's vegetable supply, one to three seasons outDiversify
Policy reflex detectionWatches for the price-control and domestic-shortage language that preceded the 2022 chicken ban, across every itemPartnerships
Single-point-of-failure registerMaintains an explicit inventory of chains with one farm, one importer or one facility, scored separately from volumeDiversify
Substitution readiness by formDistinguishes volume security from form security, so a fresh-to-frozen switch is pre-planned rather than improvisedStockpile
Local surge modellingQuantifies how much of a Malaysian shortfall local leafy vegetable and egg production can actually absorb, and over what horizonGrow Local
Hedge correlation warningFlags where the hedge shares an exposure with the risk — as local egg production does with Malaysian eggs through imported feedAll four
FS RISK › FOOD SUPPLY VULNERABILITY ASSESSMENT

Food Supply Vulnerability Assessment

The country-by-product view of where Singapore's food comes from, and how concentrated each line is. This is the assessment layer that sits between raw import data and a scored risk: it establishes exposure, and exposure is what a trigger event is scored against.

Major Source Countries by Key Food Item

Breadth of item coverage per major source country. Malaysia is the widest single block by a large margin — covering duck, eggs, vegetables, fruits, chicken, pork and fish. Breadth of coverage, not share of any one item, is the correct measure of country-level systemic exposure.

Rice Import Sources

Rice sources by share of import value. Thailand, India and Vietnam dominate, with Myanmar, Japan, the United States, Cambodia and China as the remainder. Singapore now holds rice trade agreements with Cambodia, Thailand and Vietnam, and obtained an exemption from India's 2023 non-basmati export ban — two pillars covering one item.

Country–Product Exposure Matrix

Source countryItems suppliedBreadthDominant itemOwn vulnerabilitySystemic risk to SG
MalaysiaDuck, eggs, vegetables, fruits, chicken, pork, fish7Leafy vegetables, 69%Climate, ASF, fertiliser import dependence over 60%CRITICAL
BrazilBeef, chicken, pork3Beef, 58%HPAI, long sea leg, port congestionCRITICAL
AustraliaBeef, pork, mutton, fruits, vegetables5Mutton, 90%Drought, biosecurity, export policyHIGH
ChinaVegetables, fish, fruits3Vegetables, 27%Climate, phytosanitary action, export curbsELEVATED
ThailandRice, fish, vegetables3RiceMekong hydrology, export policyELEVATED
IndonesiaFish, pork2Seafood, 17%ASF, proximity effects, export licensingELEVATED
VietnamFish, rice2Seafood, 15%Mekong hydrology, aquaculture diseaseELEVATED
United StatesFruits, beef, pork, chicken4Chicken, 8–12%HPAI, trade policyELEVATED
IndiaRice, vegetables2RiceExport bans, monsoon variabilityELEVATED
New ZealandMutton, beef, fish, fruits, pork5Mutton, 7%Low; essential supplies agreement in placeLOW
Singapore (local)Eggs, fish, vegetables3Eggs, ~30%Feed import dependence, land and sea spaceBUFFER
Local production appears in this table deliberately. It is a source like any other, with its own vulnerability profile — and its correlation with imported feed means it is not a fully independent hedge.
FS RISK › CONCENTRATION & SINGLE POINTS OF FAILURE

Concentration & Single Points of Failure

Disruptions can take place anywhere in the supply chain. A national risk register that lists only source countries misses the layers below — the single farm, the single importer, the single facility, the single corridor. Every one of these is a place where a diversified portfolio collapses to a single number.

CHAINS WITH A SINGLE FARM
1
Live pigs
CHAINS WITH A SINGLE IMPORTER
1
Live pigs
CHAINS ABOVE 60% ONE COUNTRY
3
Leafy vegetables, mutton, beef
CHAINS ABOVE 80% TOP-2
4
Chicken, eggs, beef, mutton

Single-Point-of-Failure Register

Criticality against redundancy for each identified single point. Items in the upper-left — high criticality, low redundancy — are where a single event removes a national supply line. Prototype scoring.
Single pointLayerWhat depends on itRedundancyFallbackRating
Sole live pig farm and importerSource + importerThe entire fresh pork lineNoneChilled and frozen from 20-plus sourcesCRITICAL
Sole live pig abattoirProcessingSame-day fresh pork distributionNoneImported chilled and frozenCRITICAL
Cameron Highlands growing regionSource regionA large share of imported leafy vegetablesPartialAlternate regional sources plus local surgeCRITICAL
Land crossings to MalaysiaCorridorFresh vegetables, fruit, eggs, live animalsTwo crossingsSea and air, at cost and delayHIGH
Australian mutton supplySource country90% of mutton and lamb importsNew Zealand at 7%Limited; other origins are smallHIGH
Brazilian protein complexSource countryBeef 58%, chicken 48%, pork 39%Regionalisation within BrazilMultiple accredited alternates per itemHIGH
Cold storage and powerNational nodeFrozen stockpile, commercial stock, frozen substitution surgeMultiple sites, one gridNone for a prolonged power eventELEVATED
Wholesale market and fishery portNational nodeFresh produce and seafood price formation and distributionSingle site eachDirect distribution, at reduced efficiencyELEVATED
Imported feed for local farmsUpstream inputLocal egg and aquaculture production — the Grow Local bufferMultiple feed originsAquafeed innovation under developmentELEVATED
The last row is the structural point of this register. Singapore's regenerative buffer is itself import-dependent at the input layer. Feed and fertiliser are where national food resilience is quietly rented rather than owned.
FS RISK › RISK SCORING MATRIX

FS Risk Scoring Matrix

Food security data risk mapping and modelling, using a risk-weighted score for variables. Each variable is scored, weighted by its contribution to national food insecurity, and aggregated into the risk value the kernel tests against threshold. The weights are the policy judgement; the scoring is the machine's job.

Risk-Weighted Variable Contribution

Weighted contribution of each variable class to the composite FS Risk score, shown against current observed level. Weights are set by policy and reviewed periodically; observed levels are produced by the sensing layer. Prototype weighting.

Variable Register and Weighting

Variable classWhat it measuresData sourcesWeightCurrent levelState
Source concentrationShare held by top country and top two countries per item, and country breadth across itemsImport data, accreditation registry18%HighRED
Disease and biosecurityActive outbreaks, zone extent, suspension and restriction hazardWOAH, national veterinary services16%HighRED
Trade and policyExport restrictions, quotas, price controls, pre-ban signallingGazettes, ministerial statements, WTO15%ElevatedAMBER
Climate and harvestHydrology, ENSO, rainfall anomaly, crop condition in source regionsECMWF, NOAA, crop indices14%ElevatedAMBER
Upstream inputsFeed, fertiliser, energy and freight cost and availabilityCommodity markets, chokepoint transits12%HighRED
Logistics and routesChokepoint state, transit variance, modality capacityAIS, port notices, air cargo rates10%ElevatedAMBER
Buffer positionStockpile coverage days by food group and local surge headroomStockpile records, farm production returns8%ElevatedAMBER
Demand and sentimentConsumption trend, panic-buying precursors, form substitution willingnessSearch interest, retail signals, sentiment5%NormalGREEN
Financial and counterpartySupplier and importer credit stress, non-performing exposure in the food chainFinancial data, trade credit indicators2%NormalGREEN
Weights sum to 100 and are the explicit policy layer of the system. Publishing them internally is what makes the composite score auditable — an officer can always answer why a number moved.

Scoring Method

StepOperationClaim element
1Trigger event data identified from the preliminary data set by machine learning for natural language processingClaim 1
2Trigger event validated by the verification layer before entering the registerClaim 1
3Identified trigger compared against geospatial data obtained in real timeClaim 1, characterising
4Identified trigger compared against weather forecast dataClaim 1, characterising
5Risk value calculated with positive and negative score componentsClaim 1, characterising
6Variable-class weights applied to produce the composite FS Risk scoreWeighting layer
7Impact percentile calculated and tested against thresholdClaim 1, characterising
8Classified high, intermediate or low; then supply-side or demand-sideClaims 1 and 2
9Response plan determined and activated; entity alerted via sensor-based alertingClaims 1 and 4
Steps 1 to 5 and 7 to 9 are the granted claim executed literally. Step 6 — the weighting layer — is the national policy overlay, and is where the agency's judgement enters a system that is otherwise mechanical.
INTELLIGENCE › FIEWS EARLY WARNING FRAMEWORK

FIEWS — Food Insecurity Early Warning Framework

The platform's early warning layer is built on the FIEWS framework: an early warning system for food security using a comprehensive set of indicators across the food production and consumption continuum, to give timely notice when a food crisis threatens and to elicit appropriate responses. FIEWS uses triggers derived from complexity analysis, operationalised into a real-time system that maintains food supply resilience through response plans.

SOLUTION 01

FIEWS — AI-Driven Early Warning

Continuous scanning of specific food insecurity threats, with projections generated by algorithms and the triggering of response plans to maintain supply resilience.

SOLUTION 02

Command & Control Centre

Food data intelligence assembled into a single operating picture — crisis monitoring, live data feed, predictive modelling, real-time impact assessment and prescriptive analytics.

SOLUTION 03

Productivity Platform

AI applied to food system productivity across the value chain, so the same data asset improves efficiency as well as resilience.

FIEWS Elements and Outputs

#ElementWhat it producesStatus in this build
1Complexity map of national food securityThe ontology of items, sources, routes, nodes and dependencies, with causal edgesLAYER 3
2Event triggers and indicators for insecurityThe trigger event register and the indicator set behind each oneVIEW 13
3Causal relationships and pathways between triggers and supplyThe transmission chains — feed to export ban, fertiliser to yield, rainfall to vector to import banVIEW 23
4Response plans for triggers and resilience categoriesThe pillar sequencing rules and the nine playbooksVIEWS 04 · 28
5Machine learning algorithms for risk decisioningThe six-model bank and the patented risk determination kernelVIEWS 20 · 03
6Operational system integrating all componentsThe C3 surface and the agentic staging layerVIEWS 01 · 28
7System developed and validated using historical eventsPrecedent replay against Singapore's documented disruptionsVIEW 19
The seventh element is the one most systems skip. Validating against historical events is what converts a plausible model into a defensible one, and it is why the disruption history view carries achievable lead times rather than narrative summaries.

Trigger-to-Response Pathways

Detection lead time against response lead time required, by trigger class. Classes above the diagonal can be pre-empted; classes below it can only be mitigated after the fact. Moving a class above the line is the entire purpose of the sensing investment. Prototype scoring.

Who Benefits from a FIEWS Risk Management System

StakeholderHow the collected data helps them
Food growersWeather data assists crop and stock management; market pricing data supports farm business decisions
Food processorsData on supply chain disruptions enables them to find alternative sources of supply before a line stops
Logistics providersDisruption data supports planning of alternative transport modes and routes
Exporters and importersDisruption data that will affect earnings helps them locate alternative sources in time
Food retailersDisruption data enables them to find alternative sources and stage substitution before shelf gaps appear
Banks and financing companiesData on market disruptions affecting borrowers — farmers, processors, importers — supports evaluation of non-performing loan risk
Insurance companiesData on disruption and financial loss supports risk evaluation and pricing
GovernmentDisruption data enables planning ahead to avert political crisis or instability
The last row is the reason food security is treated as a national security matter rather than a commercial one. Everything above it is why the enterprise distribution tiers exist.
INTELLIGENCE › DATA TYPOLOGY & UNIVERSE

Data Typology & the Data Universe

Managing the data universe is the unglamorous half of the platform. Without an explicit typology — what each data class is for, where it comes from, and what decision it serves — a food security system becomes a collection of feeds nobody can reason about.

Main Data Typology

Type of dataPurposeFeeds
Weather and climate changeAssists crop management; establishes the impact of weather and climatic change on crop yields in source regionsClimate model, kernel weather comparison
Production and crop yieldHistorical data on past yields to support planning and to detect deviation earlySupply model
Market pricesPrice formation as both an outcome signal and a leading indicator of policy interventionPrice and cost model
Diseases and natural disastersEvents that impact food production, processing and tradingDisease propagation model
Logistics and movement patternsPast disruption data to plan alternative modes and routes when disruption occursTransit distribution model
Export and importTrade trends by food item, country and historical patternConcentration and substitution models
Political and policy decisionsPast decisions that have disrupted supply, including statements and positions of politically exposed personsTrade policy agent, restriction model
Financial performanceCredit risk and non-performing loan data from banks and financing companies in the food chainCounterparty variable class
News and sentimentData used to anticipate political crisis or instability before it becomes policyLLM extraction, sentiment analysis

Singapore-Specific Data Typology

Type of dataPurpose in the national context
Import prices of food itemsPrice trends for the major source countries and the long tail of remaining sources. Exchange rates and supplier margins have been identified as determinants of food import prices
Retail prices, imported and localThe consumer-facing outcome, and the earliest visible sign that a supply-side event has become a demand-side one
Crop and fish yields at local farmsHistorical yield data for both traditional and urban farms, supporting planning of present and future production capacity
External political and policy decisionsPast decisions that have disrupted supply — export halts, port closures — and their measured impact on Singapore
Diseases and natural disastersAny event affecting food supply from the countries Singapore imports from
Supply chain interruptionsClosure of ports, warehouses and factories in the major source countries
Weather and climate changeSupports local farm management, and evaluates the impact on food exporters to Singapore
Financing dataCredit risk and non-performing exposure of entities in the food supply chain that rely on bank financing
Note the framing of the first row: the major source countries are tracked separately from the remaining long tail. That distinction is the analytical basis of the concentration scoring in the FS Risk section.

Data Lifecycle

Generation
UNSTRUCTURED · LOW IMPACT
Ingestion
STRUCTURED
Repository
WITH GOVERNANCE
Processing
ENRICHED
Analytics
DESCRIPTIVE · DIAGNOSTIC
Insights
PREDICTIVE
Actions
PRESCRIPTIVE · HIGH IMPACT
Impact rises left to right. Most food data systems stop at analytics, which is exactly where value stops accumulating — descriptive output has limited impact, prescriptive output drives results.

Data Value Chain in the Agri-Food System

STAGE 01

Data acquisition

Stakeholders provide quality data inputs into the collaboration platform. Generation occurs through ordinary agri-food activity, not through a separate reporting burden.

STAGE 02

Data preparation

Custom interfaces reconstruct and categorise raw input into useful, targeted data categories and control access to them.

STAGE 03

Analysis and exchange

Stakeholders acquire the data they need for their own decisions, scoped to their entitlement.

STAGE 04

Act on insights

Insights are shared back with the stakeholders who contributed, complementing the data they provided. The loop is what sustains contribution.

OPERATIONS › REGULATORY COMPLIANCE MINING

Regulatory Compliance Mining

Regulation is a data source, not a constraint on the data. Import requirements, trade circulars, gazettes, statutes and advisories are continuously mined by NLP so that eligibility is known automatically at the moment a substitution is proposed — rather than checked manually after an option has already been discussed.

DOCUMENT CLASSES MINED
9
Statutes, regulations, circulars, advisories, gazettes
SOURCE JURISDICTIONS
180+
Plus Singapore's own instruments
CHANGE DETECTION LAG
<24h
From publication to eligibility update
ELIGIBILITY CHECKS / DAY
Auto
Every proposed substitution, before it is shown

The Baseline Rule the Mining Enforces

All commercial food imports entering Singapore must come from accredited food establishments in approved countries, and only traders licensed or registered with the agency may bring in commercial shipments of food. Every substitution the decision engine proposes is tested against that rule before it appears as an option — which is why an unaccredited alternative is shown as ineligible rather than as a cheaper choice.

Mined Instrument Register

Instrument classWhat is extractedFeedsUpdate cadence
Singapore trade circularsSuspensions, liftings, affected countries and product scope, effective datesSource eligibility, corridor open-probabilityOn publication
Import requirements by itemConditions, certification requirements, permitted forms and treatmentsSubstitution feasibility modelOn change
Accreditation and establishment listsApproved countries and listed establishments per itemEligibility gate, accreditation pipelineContinuous
Source-country gazettesExport restrictions, quotas, price controls, subsidy and intervention measuresTrade policy agent, restriction modelContinuous
Veterinary and phytosanitary ordersMovement controls, zone declarations, disease-free attestationsRegionalisation zone evidenceContinuous
Food safety regulatory limitsAdditive, contaminant and residue limits, and changes to themTesting targeting, non-compliance clusteringOn change
Labelling and packaging rulesDeclaration requirements including for genetically modified foodConsignment compliance checksOn change
Novel and insect food frameworksPre-market approval status and approved listsAlternative protein supply optionsOn change
Primary legislationStatutory powers, obligations and implementation tranchesAuthority model for the agentic layerOn amendment
Trade circulars are the highest-value class. The rolling suspension and lifting cycle for avian influenza across a dozen source countries is, in effect, a live eligibility feed — and treating it as structured data rather than as correspondence is what lets the substitution set stay accurate day to day.

News and Sentiment Mining

TechniqueApplied toWhat it detects
Web crawling and scrapingNational and international media, industry press, local-language outletsEmerging disruption reporting before it reaches official channels
Natural language processingUnstructured news and official statements in the languages of source countriesStructured events with country, item, mechanism and effective date
Sentiment analysisStatements by politically exposed persons and ministriesPolicy direction ahead of formal decision — for example, signalling toward an export halt
Entity and relationship extractionCompany names, establishments, ports, growing regionsLinks a news event to a specific node in the national ontology
Risk labellingEvery ingested itemClassifies each story as stable, neutral or risky before it enters the register
Worked examples from the historical record: the announcement of Polish egg imports and Saudi Arabian shrimp imports as diversification moves; and competing international demand for shrimp supply affecting the price available to Singapore. Both are ordinary news stories, and both are supply signals once they are structured.
ENTERPRISE › FOOD VALUE CHAIN DASHBOARD

Food Value Chain Management Dashboard

The operating surface for an importer, trader or processor. Alerts, a risk-labelled newsfeed, a supplier risk matrix, importer and exporter listings, and the produce and origin picture — assembled from the same national data spine that produces the officer view, but scoped to what a firm needs to act on today.

Alerts & Reminders

RISKYSuspension notice affecting a source in your declared portfolio — poultry, two countries
RISKYFertiliser cost escalation flagged in a regional growing area supplying your vegetable line
NEUTRALNew source country approved for an item you import — eligibility now available
STABLEChokepoint transit normalised on your primary sea route
STABLELaboratory turnaround for your consignment class within normal range
NEWSFEED RISK LABELLER
RISKY NEUTRAL STABLE

Every ingested story is labelled before it reaches the user. Volume is not the product — triage is.

Supplier Risk Matrix

Supplier countryCurrent risk scoreTrendState
Country A32.50RisingRED
Country B4.80RisingAMBER
Country C2.90FlatGREEN
Country D2.00FlatGREEN
Country E1.70FallingGREEN
Country F1.20FlatGREEN
Portfolio total93.00RisingAMBER
Country labels are anonymised in this prototype. The structural point is that a single supplier country can carry a third of a firm's total portfolio risk score while representing a much smaller share of its volume.

Supplier Risk Concentration

Risk score against volume share by supplier country for a representative portfolio. The gap between the two is the firm's mispricing of its own exposure. Prototype figures.

Produce Imported by Origin

Breadth of Products by Exporting Country

Left: share of arriving goods weight by origin for a representative portfolio. Right: count of distinct product lines supplied by each major exporting country — the breadth measure that surfaces correlated country exposure at firm level exactly as it does at national level.

Importer and Exporter Listings

RoleProduct categoryWhat the listing supportsData source
ImporterAgricultural commodities — beans, pulses, nuts, oil seeds and raw intermediariesIdentifying who already holds a warm link to an alternate sourceLicence registry, declarations
ImporterCanned fruits, canned vegetables, meats and seafoodMapping the shelf-stable substitution capacity that supports the fibre stockpileLicence registry, declarations
ImporterCereals and canned foodsCarbohydrate buffer depth outside the government stockpileLicence registry, declarations
ExporterProcessed and re-export linesUnderstanding what leaves again, so net national availability is scored correctlyExport declarations
BothFresh produce and seafoodThe time-critical, least substitutable flows that need the earliest warningWholesale and port data
Re-export matters. A share of what arrives leaves again, so gross import volume overstates what is actually available domestically. Scoring net availability rather than gross arrivals is a correction the national picture needs and a per-firm view does not.
ENTERPRISE › ECOSYSTEM & DATA EXCHANGE

Ecosystem & Data Exchange

The platform connects the hard and soft dimensions of the food system — people, infrastructure and process — through data, modelling and machine learning, so that each stakeholder's strategy, programme and decisions are informed by the same picture. Nobody in a food supply chain holds enough information alone; the exchange is the point.

Who the Platform Serves

Government

Policy makers and regulators. Disruption data enables planning ahead to avert crisis, and gives every national decision an evidence chain.

Food producers & growers

Local farms and overseas producers. Weather, yield and market price data support production and business decisions.

Processors & manufacturers

Need lead time to qualify alternative ingredient origins — the stakeholder for whom early warning is worth most per week.

Importers, exporters & distributors

Hold the primary substitution decision. Disruption data directly determines earnings and continuity.

Retailers & F&B outlets

Face the demand side. Substitution staged in advance rather than explained after a shelf gap appears.

Banks, financiers & insurers

Disruption and loss data supports credit risk evaluation, non-performing loan assessment and insurance pricing across the food chain.

Logistics & input suppliers

Route, modality and input cost intelligence for planning capacity ahead of a surge rather than during one.

Trade associations, R&D bodies & non-profits

The distribution channel that reached hawkers and stallholders in past disruptions, and the research base that improves the models.

Data Inflow and Outflow

DirectionData classesContributed or consumed by
InflowFarm produce, farming operations, supplier input, distributor, transactional, commercial, logistics, market demand, consumer demand, financial, R&D, licensing, legal, climate and sentiment dataEvery ecosystem participant, plus public and administrative sources
ProcessingSupply and demand monitoring, early warning, risk scoring, inferenceThe platform itself
OutflowRisk states, substitution sets, early warnings, sector dashboards, portfolio exposure, policy analyticsGovernment, industry, associations, financiers and researchers, each scoped to entitlement

Customer Questions the Platform Answers

StakeholderTheir questionPlatform answer
Importer / exporterAre my food suppliers reliable? Is my supply chain resilient? What are my alternative sources? Can I make decent margins?Supplier risk matrix, verified substitution set, cost-of-switch estimates
GrowerWhat should I grow or rear? Will there be demand? Do I have the right conditions? How do I manage disease?Demand outlook, climate and disease intelligence, yield benchmarking
Processor / distributorHow do I lower cost of operations? How do I maximise revenue? How do I improve service quality?Input cost outlook, arrival forecasting, allocation planning
Retailer / F&BWhat do customers want? At what price will they buy? What is their purchasing behaviour?Consumption trend, form substitution willingness, demand surge signals
R&D and innovationWhat should I research? Where is the market demand? How do I commercialise? How do I do prior art and technology due diligence?Gap analysis against national targets, demand signals, IP landscape
GovernmentWhich pillar do we pull, when, and at what cost?Scored risk, pillar sequencing, staged response options
VALUE › BUSINESS MODEL & REVENUE

Business Model & Revenue Streams

The platform is structured as a licensable business system rather than a bespoke build. Licensing allows scalable deployment while preserving control over the technology and the intellectual property, aligns partner interests, and supports market penetration through controlled partner deployment rather than direct delivery everywhere.

STREAM 01

Licensing

Revenue from licensing the proprietary method and reference architecture to clients and partners, enabling authorised use and deployment.

STREAM 02

Development fees

System development and customisation to the client's own ontology, data sources and institutional workflow.

STREAM 03

Training fees

Capability transfer so officers and analysts can operate, interrogate and defend the system without vendor dependency.

STREAM 04

Platform subscription

Recurring access fees, including the enterprise tiers, providing predictable revenue and a continuous feedback loop.

Revenue Trajectory

Illustrative revenue build across the first three years, with growth driven primarily by licensing and subscription as market penetration accelerates. Figures are the planning case, not a forecast.

Market Positioning

Against traditional models

Conventional supply chain models focus on transactional services and limited risk management. This platform is proactive risk identification and mitigation, which is a different product bought by a different budget holder.

Knowledge engineering, not tooling

The differentiation is embedding expert domain knowledge into AI systems — the ontology, the weightings, the pillar sequencing rules — rather than supplying generic technology and leaving the domain work to the buyer.

Continuous improvement

The agentic framework learns from new data and user interactions, improving predictive accuracy and the quality of operational recommendations over successive cycles.

Target Markets

MarketWhy it fitsEntry routePriority
SingaporeHighest import dependence, strongest institutional capacity, and the patent's home jurisdiction. The reference implementation.National agency deploymentFIRST
ASEAN food-importing economiesShared transboundary disease, climate and route exposure; existing regional cooperation frameworks on food securityRegional cooperation, bilateralSECOND
GCC statesComparable import dependence and stockpiling posture, with additional chokepoint concentration and strong appetite for food security infrastructureGovernment-to-government, sovereign programmesSECOND
Wider Asia import-dependent economiesAcute food security risk from import reliance and complex supply chainsPartner and system integrator channelTHIRD
Enterprise and industryImporters, processors, retailers and caterers with real substitution decisions and no visibility of the whole boardSubscription, association channelPARALLEL
Financial and insuranceCredit, trade finance and insurance need scored, auditable supply chain risk as an underwriting inputData product licensingADJACENT

Deployment and Operating Model

ElementApproach
Phased command centre deploymentCommand and control centres deployed in phases by region, allowing focused resource allocation and gradual operational scaling
Strategic location selectionCentres placed where they are operationally relevant to regional supply chains, not where they are convenient to build
Local partnershipsCollaboration with local entities to strengthen infrastructure integration and operational effectiveness
Infrastructure integrationIntegration with existing regional and national infrastructure rather than parallel systems
Onboarding and trainingStructured onboarding with tailored training so users reach competence, followed by ongoing support
Cost structureContinuous development to extend the models, marketing to build the client base, and operations to maintain service quality
Singapore's role in this model is not only as first customer. It is the trusted, strategic data node in the Asian food value chain — the natural location for a regional food security command centre, and the jurisdiction whose institutional credibility makes the reference implementation persuasive elsewhere.
MY BUSINESS › IMPORTER DASHBOARD

Importer Dashboard

Representative profile: a licensed meat and fish importer holding a Licence for Import, Export and Transshipment of Meat and Fish Products, with eleven declared items across nine source countries. All figures on this page are illustrative and exist to demonstrate the operating model.

PORTFOLIO RISK SCORE
62/100
Sector benchmark 71
ITEMS AT RISK
3/11
Amber or red this week
SOURCE COUNTRIES
9
Of 43 in the accreditation database
VERIFIED ALTERNATES
17
Accredited and currently exporting
WARM LINKS HELD
5/17
Relationships that actually exist

My Portfolio Risk, Rolling 9 Months

Item-level risk for your declared portfolio. Poultry carries the steepest deterioration, driven by avian influenza activity across several of your approved sources and the northern hemisphere seasonal peak.

My Items

ItemPrimary sourceShareAlternates accreditedWarm linksDays coverRisk
Chicken, frozenBrazil61%6234HIGH
Chicken, chilledMalaysia78%316HIGH
Shell eggsMalaysia54%4211MEDIUM
Pork, frozenBrazil43%9441MEDIUM
Pork, chilledAustralia37%5322LOW
Beef, frozenBrazil52%8338MEDIUM
Beef, chilledAustralia44%6319LOW
Mutton, frozenAustralia88%2127MEDIUM
DuckMalaysia59%3114MEDIUM
Seafood, frozenVietnam31%12533LOW
Processed meatThailand35%7346LOW
Two chilled lines carry the shortest cover and the fewest warm links. Chilled product cannot be stockpiled the way frozen can, so days of cover and relationship depth are the only real defences on those rows.
MY BUSINESS › SUPPLY CHAIN RISK MAP

Where Are My Supply Chain Risks?

The first question every importer asks, answered structurally rather than anecdotally. Risk in a food import business sits in four places: the source country, the individual accredited establishment, the route, and your own concentration. Most firms track only the first.

Risk Exposure by Item and Source

Concentration against risk score for each line in your portfolio. Anything in the upper right is a line where a single event removes most of your supply and you have not built the alternative. Prototype scoring.

Four Layers of Your Exposure

LAYER 01

Source country

Disease outbreaks, export restrictions, climate events and currency. The layer everyone watches, and the only one visible in trade press.

3 countries in alert
LAYER 02

Establishment

Your actual accredited plants. A suspension can hit one establishment while the country stays open — or the reverse. Country-level news will not tell you which.

2 establishments suspended
LAYER 03

Route and modality

Chokepoints, freight capacity, cold chain integrity and clearance. A source can be perfectly healthy and still unreachable in your window.

1 route degraded
LAYER 04

Your own concentration

The layer you control. Two of your eleven lines sit above 75 per cent single-source. That is a decision, not a market condition.

2 lines above 75%

Establishment-Level Watch

EstablishmentCountryItemYour volume shareAccreditationCurrent conditionState
Est. 1042BrazilChicken, frozen38%LISTEDInside a state with active HPAI reportingRED
Est. 0871BrazilBeef, frozen29%LISTEDOutside affected zones; regionalisation holdsGREEN
Est. 2210MalaysiaChicken, chilled78%LISTEDSingle establishment carrying the whole lineRED
Est. 1730AustraliaMutton, frozen61%LISTEDStable; drought pressure on upstream flockAMBER
Est. 0459ThailandProcessed meat35%LISTEDStableGREEN
Est. 3308VietnamSeafood, frozen31%LISTEDStable; monitoring aquatic disease indicesGREEN
Est. 1188PolandPoultry, frozen0%SUSPENDEDCountry suspension for avian influenza in forceRED
Est. 2645DenmarkPork, frozen9%LISTEDStableGREEN
Establishment codes are anonymised in this prototype. In deployment they map to the actual entries in the Accreditation Database for Overseas Sources, so a suspension circular updates your risk map within the hour rather than when someone notices it.
MY BUSINESS › COST & MARGIN IMPACT

Cost & Margin Impact

A disruption is a margin event before it is a supply event. This view prices what each scenario would do to your landed cost, so the decision to hold cover, forward-contract or switch source is made on a number rather than on nerve.

LANDED COST INDEX
114
Rebased 100 = start of year
WORST-CASE SWITCH COST
+24%
Chilled chicken to the nearest alternate
BEST-CASE SWITCH
−4%
Frozen substitution on the pork line
CONTRACT REPRICING WINDOW
6wk
Lead needed to reprice customer contracts

Landed Cost Under Each Disruption Scenario

Modelled landed cost per unit under each scenario, against the base case and your contracted selling price. Where a scenario line crosses the selling price, the disruption is a loss-making event rather than a margin-compressing one — and that is the point at which repricing has to have already happened.

Switch Economics by Line

LineIf primary source failsForm changeCost deltaLead timeCustomer impactNet position
Chicken, chilledAlternate regional chilled sourceNone+24%4 wkPrice pass-through requiredMARGIN LOSS
Chicken, chilledFrozen from an accredited long-haul sourceChilled to frozen−6%6 wkCustomer acceptance neededCONDITIONAL
Chicken, frozenSecond accredited source, same formNone+11%5 wkAbsorbableMANAGEABLE
Shell eggsAlternate accredited sourceNone+16%6 wkPartial pass-throughCONDITIONAL
Mutton, frozenSecond source, thin marketNone+29%9 wkFull pass-through requiredMARGIN LOSS
Pork, frozenAny of eight accredited alternatesNone+3%4 wkNegligibleMANAGEABLE
Beef, chilledAlternate source, same gradeNone+7%5 wkAbsorbableMANAGEABLE
Note the second row. A form change from chilled to frozen is cheaper than a same-form switch — but it needs customer acceptance, which takes longer to secure than the freight does. The commercial constraint is downstream, not upstream, and it can be worked on during the warning window.
MY SOURCES › APPROVED SOURCE EXPLORER

Approved Source Explorer

All commercial food imports entering Singapore must come from accredited food establishments in approved countries, and only traders licensed or registered with SFA may bring in commercial shipments. This view mirrors the Accreditation Database for Overseas Sources so that eligibility is known before a source is considered, not after.

COUNTRIES IN ADOS SFA
43
Countries and regions with accredited establishments
APPROVED FOR HIGHER-RISK ITEMS SFA 2024
42
Livestock, meat and eggs
EU MEMBER STATES SFA 2026
18
Approved for meat and egg, up from 15
CURRENTLY SUSPENDED
Rolling
HPAI suspension list changes near-monthly

Accredited Source Countries SFA ADOS

Argentina
MEAT · POULTRY
Australia
BEEF · MUTTON · PORK
Austria
MEAT
Belgium
MEAT
Brazil
BEEF · CHICKEN · PORK
Brunei
BEEF · POULTRY
Canada
MEAT · PORK
Chile
MEAT · SEAFOOD
China
VEG · SEAFOOD · MEAT
Colombia
MEAT
Czech Republic
MEAT
Denmark
PORK · POULTRY
Finland
MEAT
France
MEAT · POULTRY
Germany
PORK · POULTRY
Hungary
POULTRY
India
RICE · VEG
Indonesia
SEAFOOD · PORK
Ireland
BEEF · PORK
Italy
MEAT
Japan
BEEF · SEAFOOD
Lithuania
EGGS · POULTRY
Malaysia
EGGS · VEG · POULTRY · PORK
Mexico
MEAT
Netherlands
MEAT · POULTRY
New Zealand
MUTTON · BEEF · SEAFOOD
Paraguay
BEEF · PORK · POULTRY
Poland
BEEF · POULTRY · EGGS
Portugal
PORK
Russia
MEAT
South Africa
FRUIT · MEAT
South Korea
BEEF
Spain
PORK · POULTRY
Sweden
MEAT · POULTRY
Switzerland
MEAT
Taiwan
SEAFOOD · PROCESSED
Thailand
CHICKEN · RICE · SEAFOOD
Türkiye
POULTRY
Ukraine
POULTRY · EGGS
United Kingdom
MEAT · POULTRY
United States
CHICKEN · BEEF · PORK · FRUIT
Uruguay
BEEF
Vietnam
SEAFOOD · RICE
Country list as published in the Accreditation Database for Overseas Sources. Item annotations are indicative of the principal categories each country supplies and are the platform's own summary. Colour indicates current disruption watch state, not accreditation status: red is an active concentration or suspension concern, amber is under watch.

Accreditation Mechanics an Importer Needs to Know

FactWhy it matters to you
Accreditation is government-to-governmentYou cannot accredit a supplier yourself. Applications are submitted through the competent authority of the exporting country, so your influence is via demand signalling, not application.
Country-level assessment comes firstSFA assesses the exporting country's animal health and food safety systems before evaluating individual establishments. A new country is a longer path than a new plant in an approved country.
Inspection visits may applySFA may inspect the exporting country and establishment before granting approval, which lengthens lead time on a new source.
Approval triggers a trade circularWhen a new accredited source is approved, SFA releases a circular informing all importers and updates ADOS. That circular is your earliest signal of a new option.
Sources change without notice to youApproved sources and accredited establishments change over time — added, removed or suspended. Checking ADOS and circulars before every consignment is a requirement, not a courtesy.
Live poultry has two accreditation levelsLive animals carry additional disease transmission risk, so live poultry sources face a more demanding accreditation regime than meat.
Under 5% meat content is differentProducts with less than five per cent meat content do not require SFA accreditation, though they remain regulated by the exporting country's competent authority.
MY SOURCES › EXPORTER DIRECTORY

Exporter Directory — Who Can Legally Ship to Singapore

Every accredited overseas establishment, by country and product category, with its current condition. This is the substitution universe: an option that is not on this list is not an option at all, regardless of price or availability.

Accredited Establishment Count by Country Group

Indicative depth of the accredited establishment estate by country group and category. Depth matters more than country count: a country with one accredited plant offers a single point of failure dressed up as diversification.

Directory by Product Category

CategoryScope per SFA definitionPrincipal approved originsDepthWatch state
PoultryChicken, turkey, duck, goose, quail, squab, guinea fowl, pheasantBrazil, USA, Thailand, Denmark, Netherlands, Poland, Ukraine, Argentina, Chile, Türkiye, Greece, Lithuania, Paraguay, Brunei, MalaysiaDeepHPAI ROLLING
BeefBovine meat and edible offalBrazil, Australia, USA, Uruguay, New Zealand, Ireland, Poland, Brunei, Greece, Paraguay, South KoreaDeepSTABLE
PorkPorcine meat, chilled, frozen and liveBrazil, Australia, Canada, Denmark, Spain, USA, China, Portugal, Greece, Paraguay; live pigs from SarawakDeep for chilled and frozen; single for liveASF WATCH
Mutton and lambOvine and caprine meatAustralia, New Zealand; special slaughter permit required per consignment for live sheep and goatThinCONCENTRATED
Table eggsFresh shell eggsMalaysia, Poland, Ukraine, Lithuania, Denmark and others; local productionModerateHPAI ROLLING
Processed egg productsSalted, preserved and processed eggBrazil, Lithuania, Paraguay and other accredited processing establishmentsModerateWATCH
SeafoodFish and fish productsMalaysia, Indonesia, Vietnam, Japan, Chile, Norway, Taiwan and othersVery deepSTABLE
Fresh fruit and vegetablesUnder the Control of Plants ActOver 100 countries and regions; Malaysia, China, Australia principalVery deepCONCENTRATED
Live oysters, frog legs, gameAmphibian and game categories in ADOSSpecific accredited establishments onlyThinSTABLE
Depth ratings are the platform's assessment, not an SFA classification. The categories and the poultry species definition follow SFA's published scope.

Live Suspension Register

SFA maintains a published list of countries and regions suspended for high pathogenicity avian influenza, and issues trade circulars as suspensions are imposed and lifted. Over 2024 to 2026 that cycle has run close to monthly across a dozen source countries.

CountryCategory affectedTypical scopePattern observed
United StatesPoultry, eggsState or county levelRepeated partial suspensions and liftings
France, Germany, NetherlandsPoultry, eggsRegionalSeasonal, concentrated in the northern winter
Poland, Hungary, Denmark, SwedenPoultry, eggsRegionalRecurrent
United Kingdom, SpainPoultry, eggsRegionalSuspensions imposed and later lifted
Japan, Canada, ArgentinaPoultry, eggsRegionalEpisodic
BrazilPoultryZone levelRegionalisation kept unaffected areas open in 2025
The operational lesson for an importer: your approved source list is not static, and the suspension cycle is seasonal and predictable in aggregate even when individual events are not. Building the alternate relationship in the northern summer costs far less than finding one in December.
FOOD CATEGORIES › MEAT, POULTRY & LIVESTOCK

Meat, Poultry & Livestock

Governed by the Wholesome Meat and Fish Act. The highest-control category in the SFA framework: importable only from approved sources and accredited establishments, with a veterinary health certificate for every consignment and the most demanding labelling requirements of any category. Chicken is used below as the worked example — every other item in this category follows the same pattern.

GOVERNING LAW SFA
Wholesome Meat
and Fish Act
Import, Export and Transhipment Rules
SOURCE RULE
Approved
sources only
Accredited establishments in approved countries
CERTIFICATE
Veterinary
health cert
Every consignment, from the competent authority
LABELLING FIELDS
8
Mandatory on every carton and packaging unit

Sub-Categories in Scope

Sub-categoryScope per SFASpecial conditionConcentration
PoultryChicken, turkey, duck, goose, quail, squab, guinea fowl, pheasantRolling HPAI suspensions; live poultry has two accreditation levelsHIGH
Beef and bovine offalBovine meat and edible offalRegionalisation arrangements available with several sourcesMEDIUM
PorkPorcine meat; chilled, frozen and liveASF watch; live pigs from a single licensed stateMEDIUM
Mutton and lambOvine and caprine meatSpecial Slaughter Permit required per consignment for live sheep and goatCRITICAL
GameEmu, kangaroo, ostrich, reindeer, venison, wild pigs, rabbitAccredited establishments only; thin estateTHIN
AmphibianCrocodile, frog legs and similarAccredited establishments onlyTHIN
Processed meat productsBak kwa, pizza with meat toppings, canned meat, animal oil and fat; over 5% meat contentLabel must carry the slaughterhouse name, designation number and slaughter dateLOW
Under 5% meat contentProducts with minor meat inclusionNo SFA accreditation required; still regulated by the exporting competent authorityLOW
Live poultry, livestock and live animalsAnimals and birds slaughtered for consumptionAnimals and Birds Act; highest disease transmission controlCRITICAL
SFA imposes the tightest controls on this category because meat carries a higher risk of pathogen contamination than processed food or vegetables. For an importer that translates into the narrowest legal substitution set of any category — which is exactly why the substitution engine matters most here.

Labelling Fields Required on Every Carton

Description of the meat product Country or region of origin Brand name, if any Processing establishment name and designation number Date of processing Slaughterhouse name, number and slaughter date — processed products Packing establishment name, number and packing date Batch number, location and canning code — canned Net weight per packaging unit and outer carton
Labelling is a substitution constraint, not an administrative afterthought. A new supplier that cannot produce compliant carton labelling cannot ship, however well accredited it is — so the platform checks label capability as part of the eligibility gate.

Electronic Sanitary Certificates

An eSPS certificate is an electronic sanitary certificate exchanged directly between competent authorities through Singapore Customs eServices, with the same validity as a paper certificate. The exchange with New Zealand began on 12 May 2023 and with Australia on 28 June 2023 for meat and meat products.

Operationally this shortens the document path on two of the most important red meat corridors, and removes the most common cause of clearance delay — incomplete or missing health certificates. When the substitution engine ranks arrivability, eSPS-enabled origins score higher for exactly this reason.
NATIONAL CHICKEN IMPORTS CNA / SFA
214,400t
2021
TOP SOURCE CNA / SFA
Brazil
48% of chicken imports
SECOND SOURCE CNA / SFA
Malaysia
34%, and nearly all fresh chicken
TOP-2 CONCENTRATION
82%
Two countries carry four fifths
APPROVED ORIGINS
15+
Countries accredited for poultry

Chicken Source Mix and How It Has Shifted

Source shares before and after the 2022 disruption, and the widened approved set since. At the time of the Malaysian export ban, Singapore was sourcing roughly 34 per cent of chicken from Malaysia, 49 per cent from Brazil and 12 per cent from the United States. Since then Türkiye, Greece, Lithuania, Paraguay and Brunei have been added as approved poultry origins.

The Fresh-Versus-Frozen Split Is the Whole Risk

FormPrincipal originWhySubstitutableExposure
Live and fresh chickenMalaysia, overwhelminglyOnly a land-adjacent source can deliver same-day slaughter freshness. Distance is the binding constraint, not accreditationNo — form changesCRITICAL
Chilled chickenMalaysia, Australia, ThailandShort shelf life restricts origin to air freight or short sea legsPartiallyELEVATED
Frozen chickenBrazil, USA, and a wide accredited setShelf life removes the distance constraint entirelyYes, widelyLOW
This is why the 2022 ban hurt despite Brazil supplying nearly half the national volume. Brazil could not replace fresh chicken with fresh chicken — it could only replace it with frozen. The response depended on hawkers, stallholders and consumers accepting a form change, supported by the Meat Traders' Association, with chilled chicken expedited from Australia and Thailand.

2022 Replayed Through the Tower

T−16 WEEKS
Global grain disruption drives regional feed costs sharply higher. Detectable in commodity data and producer commentary. Your input cost model flags margin compression across Malaysian poultry producers.
T−10 WEEKS
Domestic price pressure appears in the source country. Price control and subsidy measures are announced — the standard pre-restriction policy sequence, and the pattern the trade policy agent is trained on.
T−6 WEEKS
System issues an amber alert on your chilled chicken line. Substitution engine ranks accredited alternates. You are advised to open the Thailand and Australia relationships now, while there is no urgency and no premium.
T−2 WEEKS
Pre-ban language detected in ministry statements. Alert escalates to red. Recommended action: place a bridging order on frozen, and begin customer conversations on form substitution.
T−0
Export ban announced. Firms that acted at T−6 have product on the water. Firms that started here are competing for the same alternates at a premium.
T+8 WEEKS
Chilled chicken flows from Australia and Thailand; accreditation of new sources accelerates; consumers adapt to frozen. Impact mitigated — but the margin difference between acting at T−6 and T−0 sits entirely with the firm.
Nothing in this timeline required information an importer could not have. It required someone to be reading feed markets and foreign ministry statements against a specific chicken portfolio, every day, which is precisely what no single firm can staff and what the platform does at near-zero marginal cost.

Leading Indicators for Your Chicken Line

IndicatorWhat it precedesTypical leadSourceCurrent reading
Feed cost index in the source countryProducer margin compression, then export restriction10–16 wkCommodity markets, FXELEVATED
Domestic price control announcementsExport restriction within weeks4–8 wkGazettes, ministry statementsNORMAL
HPAI notifications and wild bird surveillanceCountry or zone suspension2–5 wkWOAH, veterinary servicesHIGH
Seasonality of northern hemisphere HPAIClustered suspensions across European and American sourcesSeasonalHistorical patternAPPROACHING PEAK
Establishment audit and suspension circularsLoss of a specific plant, not a whole countryImmediateSFA circulars, ADOS2 ACTIVE
Air and sea freight capacity on chilled lanesInability to execute a chilled substitution3–6 wkRates, capacity dataTIGHTENING
DISRUPTION › PAST DISRUPTIONS TO MY CHAIN

Past Disruptions to My Supply Chain

Filtered to the items you actually import. Each entry records what happened, how long the disruption bound your line, what the recovery required, and how much warning was available from public signals — which is the number that tells you what the platform is worth.

Disruptions Affecting My Portfolio, by Item

Duration of binding disruption against warning available, by event, for the items in your portfolio. Every point sits above the diagonal — in each case there was more warning available than the disruption lasted.

Event Register for My Items

EventMy item affectedWhat happenedBound my line forWarning availableWhat resolved it
Malaysia chicken export ban, 2022Chicken, chilledSource country banned exports to address its own domestic shortage, driven by feed cost escalation~10 weeks6–10 weeksChilled from Australia and Thailand; consumer shift to frozen; accelerated accreditation of new sources
ASF Pulau Bulan, 2023Pork, freshASF detected in a consignment; imports halted, removing a major fresh pork source overnightPermanent for that source2–5 weeksChilled and frozen from 20-plus sources; live pigs from Sarawak ramped up
HPAI Brazil, 2025Chicken, frozenBird flu detected on a Brazilian poultry farmNo binding effect3 weeksRegionalisation agreements already in place kept unaffected areas open
Rolling HPAI suspensions, 2024–26Chicken, duck, eggsNear-monthly suspension and lifting cycles across a dozen approved originsRecurring, weeks each2–5 weeksSwitching between accredited origins; the suspension list is the operative document
Russia–Ukraine feed shock, 2022Eggs, chickenGrain export disruption raised feed costs, then egg and poultry prices across the regionTwo quarters of cost8–12 weeksForward contracting and price pass-through; no volume shortfall
ASF Malaysia, 2025Pork, chilledCulling in the source country raised live pig prices sharplyCost event4 weeksFrozen substitution and alternate origins
Suez blockage, 2021Processed meatCanal blocked for six days; 400-plus vessel backlog and months of schedule disruption6–10 weeks of delayDaysRerouting and schedule absorption; the only event here with genuinely short warning
Hormuz and fertiliser, 2026All produce linesEnergy and fertiliser flows disrupted, raising production costs and lowering yields upstreamIn progress2–3 seasonsOngoing; the longest warning window of any event on this list
Seven of the eight events offered weeks or months of public warning. Only the Suez grounding was genuinely sudden. The recurring pattern is not that disruptions are unpredictable — it is that nobody was reading the signals against a specific portfolio.
DISRUPTION › PREDICTING DISRUPTION

How Can I Predict Such Disruptions?

By watching the layer above your suppliers. Disruptions to a food import line almost never begin in the food. They begin in feed, fertiliser, weather, disease surveillance and politics — and each of those has a measurable lead time before it reaches a consignment.

Warning Available by Disruption Class

Detection lead available against the time you need to execute a switch, by disruption class. Classes above the line can be pre-empted; classes below it can only be absorbed. Prototype scoring calibrated on the events in the register.

The Five Prediction Pathways

PATHWAY 01 · LONGEST LEAD

Input costs

Feed, fertiliser and energy move one to three seasons before food volumes do. The single most valuable early signal, and the least watched.

10–26 WEEKS
PATHWAY 02

Climate and harvest

Hydrology, rainfall anomaly and crop condition in your source regions. Quality degrades before volume falls, so quality reporting leads tonnage.

8–20 WEEKS
PATHWAY 03

Policy signalling

Domestic price pressure, subsidy and price control measures, and ministerial language. The reliable precursor to an export restriction.

4–10 WEEKS
PATHWAY 04

Disease surveillance

Wild bird surveillance, mortality clusters, veterinary input demand and vernacular clinical reporting, before formal notification.

2–5 WEEKS
PATHWAY 05 · SHORTEST

Route and freight

Chokepoint transit counts, war-risk premia, diversion rates and cold chain capacity. Short lead, but high certainty once it moves.

1–4 WEEKS

What the Prediction Engine Actually Does

StepOperationApplied to your portfolio
1 · SenseMultilingual ingestion across veterinary, trade policy, climate, maritime, commodity and regulatory sourcesScoped to your nine source countries and eleven items
2 · VerifyAdversarial multi-agent checking — assert, attempt to falsify, score source reliabilityYou receive verified events, not rumours. A false alarm costs you a premium you did not need to pay
3 · LocateGeospatial intersection of the event zone against establishment coordinatesTells you whether your plant is inside the zone, not just whether the country is affected
4 · QuantifyExposed volume as a share of your line, weighted by form substitutabilityA number you can take to your board, rather than a headline
5 · ScorePatented risk determination — positive and negative components, impact percentile against thresholdThe same kernel that scores the national picture, narrowed to you
6 · ThresholdThreshold modulated by your remaining days of cover on that lineThe same event escalates earlier on a line where you hold six days of cover than on one where you hold forty
7 · RecommendRanked, eligibility-checked substitution set with cost and lead timeDelivered with the alert, not requested afterwards
Step 6 is the one most importers underestimate. Two firms can face an identical disruption and face genuinely different risk, because risk is exposure divided by the time you have. Cover is the denominator.
DISRUPTION › SUPPLIER SUBSTITUTION ENGINE

If Supplier X Fails, Use Supplier Y

The engine answers one question and answers it before you ask: given that this source has just been disrupted, which accredited alternate should you switch to, at what cost, in what time, and can you legally do it. An option that fails the eligibility check is shown as ineligible rather than as a cheaper choice.

ACTIVE DISRUPTION
Chicken, chilled
Primary source scored HIGH
ALTERNATES EVALUATED
15
All accredited poultry origins
PASSED ELIGIBILITY
6
Accredited, unsuspended, exporting
RECOMMENDED
Rank 1 + 2
Split to preserve both cost and form

Substitution Ranking — Chicken, Chilled

RankAlternate sourceAccreditedSuspension statusFormLead timeCost deltaWarm linkRecommendation
1Thailand — chilledYESClearSame4 wk+18%YESActivate now. Same form, existing relationship, shortest path
2Brazil — frozenYESZone-restrictedChanged6 wk−6%YESActivate in parallel for volume. Begin customer conversation on form
3Australia — chilledYESClearSame5 wk+26%NOOpen the relationship. Proven route in the 2022 response
4Brunei — chilledYESClearSame7 wk+21%NONewly approved origin, limited volume. Worth establishing
5Paraguay — frozenYESClear9 wkChanged+14%NOLong lead. Candidate for a sourcing mission, not this event
6United States — frozenYESPartial, state levelChanged6 wk+9%YESViable but verify the specific state is unsuspended before ordering
Poland — frozenYESSUSPENDEDn/an/an/aNOIneligible. Country currently suspended for avian influenza
Non-accredited originNOn/an/an/an/aNONot an option. Flagged to SFA as an accreditation demand signal
The recommendation is a split rather than a single choice. Rank 1 preserves the product form your customers expect at a cost premium; rank 2 restores volume below base cost but needs a form conversation. Running both is usually cheaper than either alone, and the engine prices the blend.

How the Ranking Is Computed

Each candidate scored across five dimensions. A candidate must clear the eligibility gate absolutely — it is binary, not weighted — and is then ranked on the remaining four. Prototype scoring.
GATE · BINARY

Eligibility

Accredited establishment in an approved country, not currently suspended, with valid health certification available. Fail this and nothing else is computed.

FACTOR 01

Availability

Is the source actually exporting at volume today, or is it accredited on paper with no spare capacity?

FACTOR 02

Arrivability

Can it reach you inside your cover window, given route state, freight capacity and clearance? Modelled to P90, not P50.

FACTOR 03

Cost

Landed cost delta including freight, form change and any premium from other buyers competing for the same alternate.

FACTOR 04

Relationship

Do you hold a warm link? A cold relationship adds weeks that no amount of accreditation removes.

Standing Substitution Sets for Every Line

My linePrimary at riskRank 1 alternateRank 2 alternateForm fallbackSet health
Chicken, chilledRegional chilledThailand chilledAustralia chilledFrozen, wide setTHIN, 1 WARM LINK
Chicken, frozenBrazilUnited StatesThailandn/aHEALTHY
Shell eggsRegionalAccredited EU originLocal productionProcessed eggMODERATE
Pork, frozenBrazilSpainCanadan/aHEALTHY
Pork, chilledAustraliaCanadaFrozen conversionFrozen, wide setHEALTHY
Beef, frozenBrazilUruguayAustralian/aHEALTHY
Mutton, frozenAustraliaNew Zealandn/aCRITICAL, 2 SOURCES
DuckRegionalAccredited poultry originFrozenTHIN
Seafood, frozenVietnamIndonesiaChilen/aDEEP
Two rows need attention regardless of whether anything is currently wrong. Mutton has two viable origins globally and one warm link; chilled chicken has one warm link on a line with six days of cover. Those are the relationships to build this quarter, not during the next event.
ACT › ALERTS & TRADE CIRCULARS

Alerts & Trade Circulars

SFA issues a circular when approved sources are added, removed or suspended, and requires importers to check the latest circulars before every consignment. This view turns that obligation into a filtered feed: circulars matched automatically against your declared portfolio, with the ones that do not affect you suppressed.

CIRCULARS SCREENED / 30D
41
All classes
AFFECTING MY PORTFOLIO
6
Matched to my items and sources
ACTION REQUIRED
2
Substitution set changed
NOISE SUPPRESSED
85%
Circulars not relevant to you

My Feed

ReceivedTypeContentMy exposureActionLabel
T−2hSuspensionAvian influenza suspension imposed on a source in your approved set
Poultry and poultry products, regional scope
Rank 5 alternateSubstitution set updated; ranking recomputedRISKY
T−1dRisk signalFeed cost index in a source country crossed the historical intervention band
Pre-restriction pattern matched
Chicken, chilledAmber alert; open alternate relationships nowRISKY
T−3dApprovalNew source country approved for poultry
Adds a close regional origin to the accredited set
Chicken, chilledNew rank 4 alternate availableNEUTRAL
T−6dLiftingSuspension lifted on a European poultry source
Zone restrictions withdrawn
Chicken, frozenOrigin restored to the eligible setSTABLE
T−11dRequirementRevised procedures for processed meat and egg product exports
Commodity form must be indicated in the health certificate
Processed meatNotify your exporter; certificate format changesNEUTRAL
T−18dRisk signalAir cargo capacity tightening on chilled lanes
Reduces the fallback modality for chilled substitution
Chicken, chilled · Beef, chilledArrivability scores lowered on two alternatesRISKY
Every line is matched to a specific item and source in your declared portfolio. The 85 per cent of circulars that concern categories you do not import never reach you — which is what makes the six that do worth reading.

Alert Volume and Relevance

ACT › BCP BUILDER FOR SFA

Business Continuity Plan Builder

For every other importer the BCP is not compulsory, but SFA works with industry to strengthen supply networks through business continuity planning, and advises businesses to keep options open by maintaining warm links with multiple suppliers. A plan built on your own guesswork is weaker than one built on the national risk picture.

BCP SCORE
62/100
Sector benchmark 71
WEAKEST DIMENSION
Warm links
5 of 17 verified alternates
EVIDENCE ITEMS READY
24
Auto-assembled from your data
SUBMISSION STATUS
Draft
Two gaps to close before submission

Continuity Position Across Six Dimensions

Your position against the sector benchmark and the resilient threshold. The gap on warm supplier links is the most common weakness the desk surfaces: firms know their alternates exist on paper but have never transacted with them.

BCP Sections, Auto-Assembled

SectionWhat SFA is looking forEvidence the platform suppliesStatus
1 · Supply source mappingWhich sources you rely on and in what proportionDeclared portfolio with share by item and accredited establishmentREADY
2 · Risk identificationThe disruptions your sources are exposed toScored risk by item and source, with the national risk register as contextREADY
3 · Preventive strategiesWhat you do to reduce the likelihood and impact of disruptionDiversification position, cover days, forward contracting recordPARTIAL
4 · Alternative sourcesNamed alternates, verified as eligibleSubstitution sets per line, eligibility-checked against ADOS and the suspension listREADY
5 · Warm supplier linksEvidence the alternates are real relationshipsTransaction history against each alternate; gaps identifiedGAP
6 · Storage and handlingPremises approved and conditions maintainedInspection record and storage condition monitoringREADY
7 · Activation triggersWhat causes you to act, and who decidesThreshold settings per line, modulated by your days of coverREADY
8 · Response proceduresWhat you do when a trigger firesPlaybooks per line with cost and lead time attachedPARTIAL
9 · Review and testingEvidence the plan is live rather than filedHistorical replay of past events against your current planREADY
Section 9 is where the platform adds something a manually written BCP cannot. Replaying the 2022 and 2023 events against your current substitution sets shows whether the plan would actually have worked — which is a materially stronger submission than describing what you intend to do.
ACT › IMPORTER DIRECTORY & NETWORK

Importer Directory & Network

Only traders licensed or registered with SFA may bring commercial shipments of food into Singapore. This view maps that licensed community — by licence class and product category — so an importer can see the network it sits in, and so trade associations can coordinate a response across members the way they did in 2022 and 2023.

NON-RETAIL FOOD ESTABLISHMENTS SINGSTAT 2024
2,529
Processing, cold stores, slaughterhouses
FOOD SHOPS SINGSTAT 2024
23,589
Restaurants, eateries, caterers
SUPERMARKETS SINGSTAT 2024
717
Physical stores
FOOD STALLS SINGSTAT 2024
14,134
The most exposed, least resourced tier

Licence Classes an Importer Holds

Licence or registrationGoverning legislationCoversAdditional conditions
Licence for Import, Export and Transshipment of Meat and Fish ProductsWholesome Meat and Fish ActMeat, meat products, fish and fish productsImport permit per consignment via TradeNet; veterinary health certificate from the exporting authority; GIRO account
Licence to Import Fresh Fruits and VegetablesControl of Plants ActFresh producePhytosanitary requirements by origin
Registration to Import Processed Food Products and Food AppliancesSale of Food ActProcessed food, food contact articlesLabelling and declaration requirements
Licence for import of livestock and animal productsAnimals and Birds ActLive animals, table eggsFresh table eggs: storage premises inspection and an approved Business Continuity Plan before licence application or renewal
Licence for import of animal feedFeeding Stuffs ActFeed and feed ingredientsThe upstream input layer of the whole food system
Special Slaughter PermitPer consignmentLive sheep and goatRequired for each consignment
Prerequisites across all classes: a UEN registered with ACRA, the UEN registered and activated with Singapore Customs, and a GIRO account with SFA for licence and permit fees. Every consignment is declared through TradeNet with a valid import permit.

Importer Community by Category

Indicative distribution of the licensed importer community across product categories and firm size. The long tail of small importers carries a disproportionate share of the fresh, perishable, least substitutable lines — and has the least capacity to run its own risk analysis, which is the argument for the free advisory tier.

Directory Entries

TypeCompanyProduct categoriesCategory
ImporterZircon-SWIS Fine Foods Pte LtdAgricultural commodities; canned and processed linesGeneral food
ImporterWong Food and BeverageBeverage and processed foodProcessed
ImporterVezze International Pte LtdCanned fruits, canned vegetables, meats and seafoodCanned and preserved
ImporterUPC Network Pte LtdCereals and canned foodsAmbient
BothThong Huat Importers and ExportersAgricultural commodities and intermediariesCommodity
ImporterTat Hwa (1969) Pte LtdBeans, pulses, nuts and oil seedsCommodity
ImporterTat Hui FoodsProcessed and ambient food linesProcessed
ImporterTan Ciang Hak Pte LtdDry goods and commodity linesCommodity
ImporterTai Sun (Lim Kee) Food IndustriesManufactured and processed foodManufacturing
ImporterSkytime Hupkee Food Trading Pte LtdGeneral food tradingGeneral food
ExporterSingapore Fruits and Vegetables ExportFresh produce re-exportFresh produce
A partial extract for demonstration only, drawn from the importer and exporter listing in the source dataset. In deployment the directory is populated from the licence registry, and each entry is visible to the firm itself and to the agency — never to a competitor. Firm-level positions are never exposed across the network; only aggregates above a minimum count are published.

The Association Channel

PrecedentWhat the association didWhat the platform adds
Chicken export ban, 2022Hawkers and stallholders switched sources with support from the Meat Traders' AssociationSector dashboard so the association can see which members are exposed and coordinate switching before the ban, not after
ASF Pulau Bulan, 2023SFA worked with industry partners including the Meat Traders' Association to increase chilled and frozen suppliesPre-computed substitution sets across all members, with eligibility already verified
Sourcing missionsTrade delegations to prospective source countries to build new supplier relationshipsRanks candidate destinations by the concentration risk each new relationship would retire
Accreditation demandMembers request sources that are not yet approvedAggregates that demand into a ranked signal for the national accreditation pipeline
This is the loop that makes the industry portal worth building. Firms receive national intelligence; the agency receives aggregated demand signals about which sources industry actually wants accredited. Neither side can generate the other's half.
FOOD CATEGORIES › CATEGORY EXPLORER

Food Category Explorer

Every category SFA regulates, with the legislation that governs it, the source rule that constrains substitution, and the disruption profile that follows from both. Category determines almost everything about how a disruption behaves: an item importable from any country recovers in days, an item importable only from accredited establishments in approved countries can take a season.

REGULATED CATEGORIES
11
Distinct import regimes
GOVERNING STATUTES
6
Plus the Price Control Act for rice
ACCREDITATION-GATED
4
Narrowest substitution sets
OPEN-SOURCE CATEGORIES
3
Importable from any country subject to standards

Substitution Freedom by Category

Breadth of the legal source set against time to bring a new source online, by category. The lower-left quadrant is where a disruption becomes a genuine supply problem rather than a procurement inconvenience. Prototype scoring against the published regimes.

The Eleven Import Regimes

CategoryGoverning lawSource ruleCertificateSubstitution freedomRisk
Meat & Meat ProductsWholesome Meat and Fish ActApproved sources and accredited establishments onlyVeterinary health certificate, every consignmentNarrowHIGH
Eggs & Egg ProductsAnimals and Birds ActApproved sources; shell egg consignments from a single farmVeterinary health certificate; Salmonella Enteritidis-free accreditationNarrowHIGH
Fish & Fish ProductsWholesome Meat and Fish ActAny country for non-high-risk; accredited or listed sources for high-riskOnly for high-risk items and CITES speciesWideLOW
Fresh Fruits & VegetablesControl of Plants ActAny country or region, subject to pesticide residue limitsNone requiredWidestMEDIUM
Processed Food & Food Contact ArticlesSale of Food Act 1973Any country; Malaysian lines under the Enhanced Regulated Source ProgrammePermit under the Regulation of Imports and Exports RegulationsWideLOW
RicePrice Control Act and Price Control (Rice) Order 1990Licensed importers; government-designated warehousesLicence to import, re-export and wholesaleControlledMEDIUM
Live Poultry, Livestock & Live AnimalsAnimals and Birds ActAccredited farms only; two accreditation levels for live poultryVeterinary health certificateNarrowestCRITICAL
Animal FeedFeeding Stuffs ActFor food-producing animals; licensed importPer regimeModerateMEDIUM
Insects & Insect ProductsInsect Regulatory Framework; FSSA Catalogued Insect-like Species Order 2025Any country, from regulated premises; species must be cataloguedHealth certificate every consignmentEmergingLOW
Novel FoodNovel Food Regulatory FrameworkPre-market approval required before saleApproval, not certificateApproval-gatedLOW
Genetically Engineered Food & FeedGM and genome-edited crop frameworksConditions on use; labelling requirementsPer frameworkWideLOW
Risk ratings are the platform's assessment of disruption exposure, not an SFA food safety classification. Note the inversion that matters commercially: the categories with the lightest food safety burden — fresh produce, processed food — carry the widest legal source sets and therefore the fastest recovery, while the categories SFA controls most tightly are the ones where a single suspension is hardest to route around.

Category Risk Profiles

PROFILE A

Accreditation-gated

Meat, eggs, live animals, high-risk seafood. A suspension removes an entire origin instantly and a replacement takes months of government-to-government process. Cover and warm links are the only defences.

PROFILE B

Standards-gated

Fresh produce, processed food. Legally importable from anywhere, so the constraint is commercial and logistical rather than regulatory. Recovery measured in weeks.

PROFILE C

Licence-controlled

Rice. Not a food safety regime at all but a price control regime under MTI, with licensed importers and designated warehouses. Disruption behaves differently and is managed differently.

PROFILE D

Approval-gated emerging

Novel food, insects, genetically engineered lines. Small volumes today, but the approval pathway is the constraint on how fast alternative proteins could ever become a resilience option.

FOOD CATEGORIES › FISH & SEAFOOD

Fish & Fish Products

The category with the widest source freedom and the most intricate exception list. Fish and fish products that are not high-risk and not under CITES can be imported from any country or region without a health certificate — but a specific set of high-risk items carries conditions ranging from certification to outright prohibition.

BASE RULE
Any country
For non-high-risk, non-CITES products
PROHIBITED ITEMS
4
Chilled shucked oysters, cockle meat, cooked prawn, crab meat
LIVE OYSTER ORIGINS
10
Countries with a recognised shellfish sanitation programme
CITES SPECIES GROUPS
14+
Appendix II and III, permits from NParks

High-Risk Item Register

ProductRequirementEffect on my substitution setState
Chilled shucked raw oysters · chilled cockle meat · chilled cooked prawn or shrimp · chilled crab meatImport is not allowedNo substitution exists. Frozen equivalents are the only routePROHIBITED
Live cocklesOnly food processing establishments licensed by SFA to shuck cockles may importYour licence class determines eligibility, not your supplierRESTRICTED
Live oystersOnly from Australia, Canada, France, Ireland, Japan (Fukuoka, Hiroshima, Hokkaido, Mie, Miyagi, Oita, Tokushima, Iwate and Shimane Prefectures), the Netherlands, New Zealand, the United Kingdom and the USA. Health certificate requiredTen origins; Japan is prefecture-restricted so a single prefecture event narrows it furtherLISTED ORIGINS
Frozen oystersHealth certificate certifying the consignment has been tested and found free of norovirusAny origin able to certify; testing capability is the real constraintCERTIFIED
Frozen blood cockle meat · frozen cooked prawns · frozen raw or cooked crab meatHealth certificate from the relevant authority of the exporting countryWide origin set; certificate capability is the filterCERTIFIED
Frozen pearl oyster meat, adductor muscleAny country with documentary proof of regulated harvesting and processing, hygienic conditions and compliance with the Food Regulations; records retained for inspectionWidest of the high-risk itemsDOCUMENTED
PufferfishWild and farmed muscle meat and farmed parts — fins, skin and milt — only from SFA-accredited pufferfish establishments. Health certificate requiredEstablishment-level accreditation; very thin setACCREDITED ONLY
Seafood from FijiOnly sharkfin and sea cucumber allowed; other seafood not permittedOrigin effectively closed for general seafoodRESTRICTED
Seafood from Chile and MexicoHealth certificate certifying products are free from marine biotoxinsTwo significant origins with an added certification stepCERTIFIED

CITES-Listed Species in the Food Chain

Permits are required for importing fish species under CITES Appendix II and Appendix III, applying to whole fish, fish parts and fish derivatives. Each consignment requires both a CITES import permit from NParks and a CITES export or re-export permit from the country of export.

Sturgeon (Acipenseriformes)Whale sharkBasking shark Seahorses (Hippocampus)Thresher sharks (Alopias)Humphead wrasse Mediterranean date musselMackerel sharksHammerhead sharks (Sphyrnidae) Requiem sharks (Carcharhinidae)Sea cucumbers (Thelenota)Teatfishes Mobulid rays (Manta)European eelRock sea cucumber, Ecuador — Appendix III
The platform screens species names against CITES listings at the point a substitution is proposed, because the most common failure here is a commercially sensible switch into a species that then requires a permit chain nobody budgeted time for.

Seafood Source Diversity — Why This Category Is Resilient

Seafood carries the lowest top-2 concentration of any major category at around 42 per cent, spread across Malaysia, Indonesia, Vietnam and a long tail. Combined with the open base rule, this is structurally the most substitutable protein in the national portfolio — a useful fact when other protein lines are constrained.
FOOD CATEGORIES › EGGS & EGG PRODUCTS

Eggs & Egg Products

Shell eggs, salted eggs, liquid eggs and omelettes. The category with the tightest consignment-level rule in the entire framework, the only category with a mandatory business continuity plan attached to the licence, and one of the two most concentrated lines in Singapore's import portfolio.

CONSIGNMENT RULE SFA
Single farm
per consignment
Shell eggs may not be mixed across farms
ACCREDITATION CRITERION SFA
Salmonella
Enteritidis free
Farm-level requirement
LICENCE CONDITION SFA
Mandatory
BCP
Approved before licence application and renewal
PREMISES
Inspected
and approved
Egg storage premises, before licensing

Why the Single-Farm Rule Changes Your Risk Arithmetic

The ruleEach shell egg consignment must be derived from a single farm. You cannot blend output from several farms into one shipment to spread risk.
The consequenceDiversification has to happen at the consignment schedule level, not within a consignment. Three farms means three shipment streams, not one blended shipment.
The failure modeAn importer running a single farm across all consignments is fully exposed to one farm's Salmonella status, one farm's flock health, and one farm's HPAI zone — even though the country looks diversified on paper.
The platform's responseFarm-level rather than country-level exposure tracking for the egg line, and a substitution set built from accredited farms rather than accredited countries.

National Egg Diversification — the Measurable Case

Malaysia accounted for 72 per cent of Singapore's egg imports in 2019 and 52 per cent by 2021, while locally produced eggs rose from 26 to 30 per cent. Singapore received its first batch of Polish eggs in June 2020 and has since added Lithuania and other accredited origins. This is the clearest published evidence that deliberate diversification works — and the benchmark against which an individual importer's own concentration should be judged.

Sub-Category Requirements

Sub-categorySource ruleCertificateLabellingSubstitution
Shell eggsApproved sources; single farm per consignmentVeterinary health certificatePer veterinary conditions for importation of table eggsNarrowest
Pasteurised shell eggsApproved sourcesVeterinary health certificatePer veterinary conditions for pasteurised shell eggsNarrow
Processed and salted eggsSFA-approved sources and accredited processing establishmentsHealth certificateStandard food labelling requirementsModerate
Liquid eggs and omelettesApproved sourcesHealth certificateStandard food labelling requirementsModerate
Processed and liquid egg lines are a genuine resilience instrument. They carry a wider accredited set than shell eggs and no single-farm rule, so a shell egg disruption can be partly absorbed by shifting industrial and food service demand into processed formats — which is why recent approvals for processed egg sources matter more than their volume suggests.
FOOD CATEGORIES › FRESH FRUITS & VEGETABLES

Fresh Fruits & Vegetables

Governed by the Control of Plants Rules. Legally the freest category in the framework — importable from any country or region — and yet in practice one of the most concentrated, because perishability, not regulation, is the binding constraint. This gap between legal freedom and practical concentration is the defining characteristic of the produce line.

SOURCE RULE SFA
Any country
or region
No accreditation gate
ACTUAL CONCENTRATION CNA / SFA
69%
Leafy vegetables from one country
SOURCE COUNTRIES SFA
100+
For vegetables overall
BINDING CONSTRAINT
Shelf life,
not law
Distance is the real gate

Legal Freedom Against Practical Concentration

Number of legally available origins against actual realised concentration, by produce line. Leafy vegetables sit at the extreme: over a hundred legal origins, and roughly seven in ten kilos from one country. Hard produce and fruit sit much closer to their legal freedom because shelf life permits distance.

Import Requirements

RequirementDetailSubstitution implication
Prohibited pesticidesImports must not contain any pesticide outside the SFA-approved list, checkable through the Pesticide Search ToolA new origin's pesticide regime must be verified before first shipment, not after
Residue limitsPesticide or toxic chemical residue must not exceed the levels in the Ninth Schedule of the Food Regulations or those recommended by the Codex Alimentarius CommissionObtain laboratory test results from the supplier as part of qualification
Labelling per containerEvery carton or basket must carry the producer name and address, product description, and date of export or packingThree fields only — far lighter than meat, which is part of why produce switches faster
Reclassification on processingFruits and vegetables that have been cut, peeled, canned or frozen are classified as Processed Food and follow that regime insteadProcessing changes the applicable rules entirely, including for minimally processed lines from Malaysia

Why Produce Disrupts Differently

Quality fails before volume

In the 2022 regional rainfall event, vegetables reaching Singapore were often of poorer quality and spoiled faster, before tonnage fell. An importer watching only arrival volume sees the disruption late.

Upstream inputs bite hardest

Fertiliser cost shocks in the source region raise production costs and reduce planted area one to three seasons out. It is the earliest available signal on this line and the least watched.

Local production is the real buffer

Locally grown vegetables cushioned the 2022 shortfall. On this category — and only on this one — the Grow Local pillar is a genuine short-horizon substitute rather than a long-term ambition.

FOOD CATEGORIES › RICE & PRICE-CONTROLLED ITEMS

Rice — a Price Control Regime

Rice sits outside the food safety framework entirely. It is a controlled item under the Price Control Act and the Price Control (Rice) Order 1990 issued by the Minister for Trade and Industry. A licence is required to import into Singapore or the Free Trade Zone, to re-export, and to carry out wholesale dealings — and government-designated warehouses form part of the regime.

GOVERNING LAW SFA
Price Control
Act
Rice Order 1990, issued by MTI
LICENCE SCOPE
Import ·
Re-export ·
Wholesale
All three require licensing
STORAGE
Designated
warehouses
Government-designated under the scheme
CONCENTRATION
Region
Mekong-weighted, correlated across origins

Rice Source Structure

Rice import sources by share of value, with the Mekong-region grouping shown separately. The critical structural fact for an importer: Thailand, Vietnam, Cambodia and Myanmar are not four independent origins. They share a river basin, a monsoon and increasingly a hydrology problem, so a regional climate event correlates them.

The Two Disruption Modes on Rice

ModeMechanismDocumented instanceWhat resolves itMy lead time
Export restrictionAn exporting government restricts outbound shipments to protect domestic supply or pricesIndia banned non-basmati rice exports in July 2023Diplomatic exemption plus variety and origin switching. Singapore obtained an exemption and importers increased imports of different varieties from multiple sources3–8 weeks
Region-wide climateHydrology and monsoon failure across the producing basin reduces harvests, which take months or years to recoverOngoing Mekong hydrology deterioration; El Niño cycles spanning multiple growing seasonsDistant origins at cost, plus the national stockpile. Substitution fails when the whole region is affected8–20 weeks
Rice is the one category where the diplomatic instrument outranks commercial substitution, because the constraint is political rather than physical. Singapore holds rice trade agreements with Cambodia, Thailand and Vietnam — an importer's risk position on this line is partly determined by government-to-government arrangements it does not control but should track.

Variety Substitution

VarietyPrincipal originsSubstitutable withConsumer acceptanceRisk
Non-basmati long grainIndia, Thailand, VietnamThai and Vietnamese long grainHighMEDIUM
BasmatiIndia, PakistanLimited; variety-specific demandLowHIGH
Jasmine and fragrantThailand, Vietnam, CambodiaWithin the Mekong group onlyMediumMEDIUM
Japonica and short grainJapan, Republic of Korea, Australia, USAAustralian and US japonicaMediumLOW
Parboiled and specialtyIndia, ThailandOrigin-flexibleHighLOW
Brown and wholegrainMultiple, including Australia and USAWideHighLOW
Basmati is the trap on this line. It looks like rice in a stock report but behaves like a single-origin specialty product, with two producing countries and consumers who will not accept a substitute. It carries higher effective risk than its volume implies.
FOOD CATEGORIES › PROCESSED FOOD & ERSP

Processed Food & Food Contact Articles

Governed by the Sale of Food Act 1973 and the Food Regulations. Processed food covers products such as biscuits, cooking oil and mineral water; a food contact article is an item designed for use with food that touches it and needs no electricity, such as plates, bowls, cups, straws, pots, chopsticks and spoons. Importable from any country, with a valid permit — but with one significant origin-specific programme attached.

SOURCE RULE
Any country
With a valid import permit
GOVERNING LAW SFA
Sale of Food
Act 1973
And the Food Regulations
ERSP PRODUCT TYPES SFA
6
Malaysian lines requiring registered establishments
LEAD TIME TO REGISTER
1mo+
To gather ERSP documentary proof

Enhanced Regulated Source Programme — Imports from Malaysia

Processed food imported from Malaysia under this programme must come from establishments registered with the Malaysian Ministry of Health, with documentary proof submitted to SFA before an import permit can be applied for. On approval, an establishment code is issued for each manufacturer and must be declared in the CA/SC Code field in TradeNet.

#Product typeNotesSubstitution difficulty
1Infant formulaNutritional composition must fall within prescribed limitsHIGH
2Infant cerealVoluntary Supplier Registration scheme exempts Aflatoxin B1 testing per consignmentHIGH
3Traditional cakes and kueh, cooked rice products and other perishable ready-to-eat cooked foodFrozen and uncooked products currently excludedMEDIUM
4MooncakesHighly seasonal; a registration lapse before the festival is a total lossMEDIUM
5Dried beancurd sticks, skin and sheetsStaple ingredient lineLOW
6Edible oilCommodity line with wide alternate origins outside MalaysiaLOW
Ready-to-eat minimally processed cut or peeled fruits including coconuts, vegetables and cut sugarcane from Malaysia follow the same registration process. The operational point for an importer: gathering the MOH registration letter and invoices can take a month or longer, so ERSP registration is a task for the planning calendar, not the crisis response.

Registration Pathways

PathwayApplies toWhat you submitCode issuedTurnaround
Enhanced Regulated Source ProgrammeSix processed food types from MalaysiaMOH Malaysia registration letter and establishment invoices, via the SFA feedback formEstablishment code, e.g. MYZNNV0001Document gathering, one month or more
Supplier Registration for Infant Cereal, voluntaryInfant cereal suppliers — manufacturer, producer or consolidatorSupplier verification checklist and registration form; no feeEstablishment code, e.g. AUZGxxxxWithin two weeks
Overseas insect processing establishmentInsect and insect products for human consumptionRegulated source documents and FSMS or HACCP plan; one-time online registrationEstablishment code for TradeNetPer assessment
Specific food product documentationInfant formula, ready-to-eat food and other listed productsAdditional supporting documents per the published listn/aPer product
A rejected permit declaration is usually a code problem, not a product problem: an omitted establishment code, an invalid code, or a missing health certificate will cause TradeNet to reject the declaration outright. The platform holds your codes against each supplier so the declaration is right the first time.

Regulatory Compliance Surface

Documentary burden against source freedom by processed food line. Most processed food is the easiest category in the framework to substitute; the ERSP and infant nutrition lines are the exceptions, and they sit at the high-burden end because of who consumes them.
FOOD CATEGORIES › NOVEL FOOD, INSECTS & FEED

Novel Food, Insects & Animal Feed

The three categories that determine whether Singapore's protein options widen over the next decade. Small volumes today, but each carries an approval pathway that sets the ceiling on how quickly an alternative source could ever become a resilience instrument.

NOVEL FOOD
Pre-market
approval
Required before sale in Singapore
INSECTS SFA 2025
Catalogued
species order
FSSA Catalogued Insect-like Species Order 2025
ANIMAL FEED
Feeding
Stuffs Act
For food-producing animals
STRATEGIC ROLE
Upstream
and future
Feed constrains today; novel food widens tomorrow

Insects & Insect Products — Four Import Streams

StreamPurposeOne-time submissionPer-consignment certificationAdditional
AInsects and insect products for human consumptionDocumentary proof of regulated premises; HACCP or FSMS planHealth certificateSpecies must be in the Catalogued Insect-like Species Order 2025; declaration required for permit approval
BInsects and insect products as animal feedDocumentary proof of regulated premises; HACCP or FSMS planHealth certificate
CLive insects for local production, human consumptionDocumentary proof insects are farmed in regulated premisesHealth certificateNParks approval for import conditions for insects and invertebrates; catalogued species
DLive insects for local production, animal feedDocumentary proof insects are farmed in regulated premisesHealth certificateNParks approval required
Every consignment must certify that the insects were not harvested from the wild; that manure, decomposing organic material and material from diseased animals or fish were not used as feeding substrate; that no ruminant material was used except deboned meat cuts and dairy; and that products were handled and packed hygienically and are free from contaminants. Substrates that have undergone controlled fermentation are permitted.

Animal Feed — the Upstream Layer

Feed is not a consumer category, but it determines the cost and availability of every locally produced egg and farmed fish, and it sets the margin position of every regional poultry, egg and pork producer Singapore imports from. It is the earliest signal available on several downstream lines.

Why it matters to an importerMechanismObserved instance
It precedes export restrictionsFeed cost escalation compresses regional producer margins, which drives domestic price pressure and then trade actionGrain disruption in 2022 raised chicken feed costs, then egg prices, and contributed to export bans in the months that followed
It constrains the local bufferLocal egg and aquaculture production depends on imported feed, so a feed shock degrades the hedge at the same time as the exposureShared feed dependency between Singapore and regional producers
Innovation changes the ceilingAquafeed capability determines how far local seafood production can scale toward the protein targetA feed and feed ingredient technology centre announced in 2026 to drive aquafeed innovation for tropical marine species

Emerging Category Trajectory

Approval-pathway maturity against potential resilience contribution for each emerging category. None of these is a near-term substitution option in a disruption; all of them change the shape of the substitution set over a five to ten year horizon, which is why they belong on an importer's strategic radar rather than its operational one.

Genetically Engineered Food and Feed

FrameworkScopeImporter obligation
Genome edited crops in food and animal feedRegulatory framework for the use of genome edited cropsConfirm the framework position of any new crop origin before qualification
Genetically modified crops in foodConditions related to their useVerify compliance conditions apply to the product as formulated
GM food labellingLabelling requirements for genetically modified foodEnsure supplier labelling meets the requirement before shipment
This framework rarely constrains an existing line, but it is a live consideration whenever a substitution moves into a new origin with different agricultural technology — which is precisely what a disruption forces an importer to do at short notice.
ASEAN COMMAND › EXECUTIVE C3

ASEAN Collective Food Security Command & Coordination

Independent prototype. A concept demonstrator built on granted Singapore patent SG 10202109991Y. It is not an ASEAN system and carries no ASEAN, ASEANstats, AFSIS, APTERR or Member State endorsement. Figures tagged ASEAN, ASEANstats, APTERR or FAO are drawn from published sources; every other figure is tagged ILLUSTRATIVE and exists only to demonstrate the operating model. The platform is an analytical and decision-intelligence layer complementing existing ASEAN institutions — it does not replace their authority, and it implies no centralised control over Member States.
MEMBER STATES ASEAN
11
Timor-Leste admitted 26 October 2025
AGRI, FORESTRY & FISHERIES TRADE ASEANstats
US$447bn
2022; about 12.4% of ASEAN merchandise trade
PADDY PRODUCTION ASEANstats
195.5Mt
2022; five states produce about 90.9%
STATES BELOW THE REGIONAL AFSIM SCORE DERIVED
4
Computed by the AFSIM country layer — see Member State Comparison
REGIONAL SUBSTITUTION HEADROOM MODELLED
Thin
On rice, under simultaneous demand — AFSIM indicator E4

The Two Central Propositions

PROPOSITION 01

ASEAN food security is a network problem

A Member State may diversify nationally while ASEAN collectively remains exposed to the same producer, the same climatic system, the same river basin, the same animal disease, the same feed input, the same fertiliser source or the same shipping corridor.

The region therefore needs visibility not simply into food trade, but into correlated regional dependency.

PROPOSITION 02

National optimisation may produce regional failure

During a disruption several Member States may simultaneously switch to the same alternative supplier. Each recommendation is individually rational. Collectively they may be impossible.

The system must detect simultaneous substitution risk before governments and commercial buyers compete for an insufficient substitute supply.

Regional Supply Assurance by Commodity

Modelled probability that normal regional supply can be maintained for each commodity group under current conditions, and the collective substitution headroom behind it. Prototype series. Note the divergence: assurance can look adequate while headroom is thin, because headroom is what remains after domestic requirements and existing commitments are removed.

Regional Risk Posture by Domain

Coordination Priorities

Quantify rice substitution headroom
Four import-dependent states share the same alternate supplier set
P1
Map the shared feed dependency
Several poultry and egg producers depend on the same imported maize and soybean meal
P1
Mekong basin correlation watch
Five states, one hydrology; country diversification does not diversify climate
P1
Fertiliser import exposure
Upstream shock reaches crop yields one to three seasons out
P2
Reserve readiness assessment
Regional reserve mechanisms against modelled emergency requirement
P2
Corridor resilience review
Land crossings, archipelagic logistics and the Malacca corridor
P3

From National Reference to Regional Layer

LEVEL 01

Singapore National Food Security AIC³

The reference implementation. One import-dependent state, 180-plus source countries, eleven food categories, a working kernel and a documented disruption history. Proves the method.

39 government views19 industry views
LEVEL 02

ASEAN Collective Food Security AIC³

Regional interdependency and collective resilience. Eleven Member States treated as one interconnected food system while national sovereignty is preserved. Singapore serves as the technology, analytics, logistics, finance and decision-intelligence hub from which the capability is demonstrated.

33 regional viewsSame kernelAFSIM measurement layer

Member State Status Board

Member StateStructural rolePrincipal strengthsPrincipal dependenciesSystemic importanceState
ThailandMajor surplus and export hubRice, cassava, sugar, poultry, processed food, fruitMekong hydrology, imported feed and fertiliserCRITICALGREEN
Viet NamMajor surplus and export hubRice, seafood, coffee, fruit and vegetables, fisheriesMekong delta hydrology and salinity, aquaculture diseaseCRITICALAMBER
IndonesiaHuge producer and huge domestic consumerRice, maize, palm oil, fisheriesDomestic demand limits exportability; inter-island logisticsCRITICALAMBER
MalaysiaMixed producer and importer; critical Singapore supply hubPalm oil, poultry and eggs, vegetables, fruit, processed foodFeed and fertiliser import dependence; ASF exposureCRITICALRED
PhilippinesLarge consumption and import market with significant domestic productionBananas, coconuts, fisheriesRice and feed grain vulnerability; typhoon exposure; archipelagic logisticsHIGHAMBER
MyanmarPotential food surplus producerRice, pulses, maizePolitical and logistics risk affecting reliabilityHIGHRED
CambodiaAgricultural surplus potentialRice, cassava, maizeMekong hydrology; milling and logistics capacityMEDIUMGREEN
Lao PDRSmaller producer and land-linked corridorRice, maize, livestockLandlocked logistics; Mekong dependencyLOWAMBER
SingaporeHighly import-dependent demand and logistics hubRegional sourcing, finance, logistics, cold chain, trading capabilityOver 90% of food imported; land corridor to MalaysiaHIGHAMBER
Brunei DarussalamSmall import-dependent consumption marketFiscal capacity to hold reservesHigh reliance on external supply; small volumes limit buying powerLOWGREEN
Timor-LesteEmerging food-security vulnerability nodeNewest Member State; regional integration under waySmaller productive base, weaker trade integration, maritime supply dependencyLOWRED
Structural roles are analytical characterisations, not ASEAN classifications. State ratings are prototype kernel outputs. Every Member State plays a structurally different role — the platform must not assume otherwise, and the collective substitution engine depends on that asymmetry.
ASEAN COMMAND › FOOD INTERDEPENDENCY MAP

ASEAN Food Interdependency Map

Eleven country nodes and the food flows between them. Node size indicates systemic importance to ASEAN food security. Line width indicates relative trade volume, and line colour indicates dependency risk. The map exists to answer one question at a glance: who depends on whom, and how badly.

RICE DEPENDENCY NETWORK ILLUSTRATIVE TOPOLOGY
ThailandSURPLUS HUB Viet NamSURPLUS HUB CambodiaSURPLUS MyanmarRELIABILITY Lao PDRCORRIDOR PhilippinesDEFICIT MalaysiaMIXED / DEFICIT SingaporeFULL IMPORT BruneiIMPORT IndonesiaSWING NODE Timor-LesteVULNERABLE NODE SIZE = SYSTEMIC IMPORTANCE · LINE WIDTH = RELATIVE VOLUME · LINE COLOUR = DEPENDENCY RISK · DASHED = INTERMITTENT OR POLICY-CONTINGENT FLOW

What a Country Node Exposes

FieldWhy it is on the nodeWorked value — Philippines, rice ILLUSTRATIVE
Domestic productionThe base against which import need is calculatedLarge, but structurally below consumption
Domestic consumptionDetermines the deficit that must be met from trade or reservesHighest per-capita staple demand in the group
Imports and exportsActual flows, not intentionsPersistent net importer
Exportable surplusProduction minus domestic requirement, reserves and commitmentsNone
Top ASEAN suppliersConcentration inside the regionViet Nam and Thailand dominant
Top external suppliersWhere the region turns when internal surplus failsIndia, Pakistan
Critical upstream inputsThe dependencies behind the dependencyFertiliser, feed grain, fuel
Substitution headroomWhat is actually available after everyone else has taken theirsThin under simultaneous demand
Current kernel riskThe authoritative risk value from SG 10202109991YINTERMEDIATE

Two Sub-Systems, Three Connectors

SUB-SYSTEM 01

Mainland ASEAN

Thailand, Cambodia, Lao PDR, Viet Nam and Myanmar, linked through Mekong climate, rice, cassava, maize, cross-border trucking, river systems, animal disease, labour and trade policy.

A Mekong drought is not a Viet Nam event or a Thailand event. It is an ASEAN correlated regional event.
SUB-SYSTEM 02

Maritime ASEAN

Indonesia, Malaysia, Singapore, the Philippines, Brunei and Timor-Leste, linked through ports, shipping routes, feed imports, palm oil, seafood, processed food, cold chain and the Malacca and Singapore waterways.

Route and modality risk dominates here, where climate correlation dominates the mainland.
CONNECTORS

Thailand · Malaysia · Singapore

The three economies that bridge the mainland and maritime food systems. Their resilience matters disproportionately, because a failure at a connector propagates across both sub-systems rather than within one.

This is why connector states carry higher systemic importance than their production volume alone would suggest.
ASEAN COMMAND › INTRA-ASEAN TRADE MATRIX

Intra-ASEAN Trade Matrix

An eleven by eleven bilateral dependency matrix. Rows are exporters, columns are importers, and each cell scores the importing state's dependency on that exporter for the selected commodity. The purpose is to let a decision maker see immediately who supplies whom, and where a single relationship carries a whole national requirement.

RICE — IMPORTER DEPENDENCY ON EXPORTER ILLUSTRATIVE PROTOTYPE VALUES
DEPENDENCY none critical
Cells show the share of the importing state's rice import requirement met by that exporter, scored 0 to 9. In deployment the matrix is driven by a bilateral trade cube refreshed from ASEANstats, AFSIS, FAOSTAT detailed trade matrix, UN Comtrade and national customs data — filterable by commodity, and switchable between value, volume, share of importer requirement, share of exporter exports and dependency score.

What the Matrix Reveals That a Country Report Does Not

ReadingObservationConsequence
Read a columnAn importer's full supplier set and its concentrationIdentifies which states are single-supplier exposed on a staple
Read a rowAn exporter's full customer setIdentifies which exporters carry systemic weight — a restriction by them is a regional event
Read a column groupSeveral importers sharing the same dominant exporterThis is the simultaneous substitution signal. Four deficit states sharing two surplus suppliers is a structural fragility, not a diversified market
Compare across commoditiesThe same exporter appearing dominant in several matricesCorrelated exposure at country level — one policy decision affects several commodities at once
Read the diagonal neighbourhoodIntra-mainland and intra-maritime clusteringShows where regional trade is genuinely integrated and where it is thin

Dependency Concentration by Importing State

Top-supplier share and top-three share for staple imports by Member State. States on the right of this chart hold the least room to manoeuvre when their dominant supplier restricts exports — and they are precisely the states most likely to compete for the same alternative.
DEPENDENCY › COMMODITY NETWORKS

Commodity Dependency Networks

Each commodity forms a different network with different production nodes, different deficit markets, different corridors and different recovery times. Treating food as one aggregate hides all of it. Six networks carry most of ASEAN's collective food security exposure.

NETWORK 01

Rice

Surplus nodes: Thailand, Viet Nam, Cambodia, Myanmar.
Deficit nodes: Philippines, Malaysia, Singapore, Brunei, Timor-Leste; Indonesia swings between producer and importer with the harvest.

Recovery: months to yearsReserve instrument existsMekong-correlated
NETWORK 02

Poultry & Eggs

Supply nodes: Thailand for poultry and processed protein; Malaysia for eggs and poultry into Singapore and regional markets.
Shared exposure: imported maize and soybean meal across every producer.

Recovery: flock rebuild, monthsFeed-correlatedHPAI seasonal
NETWORK 03

Seafood

Supply nodes: Indonesia, Viet Nam, Thailand, Malaysia, Philippines.
Character: the most fragmented and therefore most substitutable protein network in the region.

Recovery: weeksCold chain dependentAquafeed exposure
NETWORK 04

Vegetables & Fruit

Character: short substitution windows, strong cold-chain dependency, border-clearance sensitivity, sanitary and phytosanitary restriction, high spoilage.

Recovery: weeksCorridor-criticalQuality fails before volume
NETWORK 05

Palm & Cooking Oil

Supply nodes: Indonesia and Malaysia, which together with the wider top-two global position accounted for around 85% of global palm oil exports in 2024 FAO.

This is production concentration risk rather than import dependency risk. ASEAN is collectively self-sufficient and simultaneously systemically exposed, because most surplus originates in one climatic geography and two neighbouring countries.
NETWORK 06

Maize & Feed

Supply nodes: Indonesia produced around 52.4% of ASEAN maize in 2022 ASEANstats; regional maize is largely consumed domestically as feed.

The most important network in the set, because it sits upstream of poultry, eggs, pork and aquaculture simultaneously. A feed shock is not one commodity failing — it is four protein networks degrading together.

Network Character Comparison

Substitutability, recovery time and regional concentration by commodity network. The pattern that drives response sequencing: seafood substitutes easily and recovers fast; rice substitutes poorly and recovers slowly; feed substitutes moderately but propagates into four other networks.

Production Concentration — the Structural Fact

CommodityASEAN production, 2022 ASEANstatsConcentrationImplication for collective resilience
Paddy195.5 million tonnesIndonesia, Viet Nam, Thailand, Myanmar and the Philippines account for about 90.9%Five producers carry the regional staple. Two of them are also large domestic consumers
Cassava85.3 million tonnesThailand, Indonesia, Cambodia and Viet Nam account for about 77%Feed, starch and industrial input concentration on the mainland
Maize44.7 million tonnesIndonesia alone produces about 52.4%The feed input for the whole regional protein system sits largely in one country, mostly consumed domestically
These three lines are the analytical heart of the ASEAN layer. Production sufficiency is genuine; the resilience question is where the food is, who needs it, through which routes it must travel, and how much is actually available to move.
DEPENDENCY › COUNTRY PROFILES

Country Food Security Profiles

Eleven Member States playing six structurally different roles. A regional platform that treats them as interchangeable will produce recommendations that are individually plausible and collectively wrong.

ROLE 01

Major surplus and export hub

Thailand · Viet Nam. Carry regional substitution capacity. Their export policy is a regional variable, not a national one.

ROLE 02

Huge producer, huge domestic market

Indonesia. Production volume does not translate into exportable surplus. Swings between exporter and importer with the harvest.

ROLE 03

Mixed producer and importer

Malaysia · Philippines. Both produce meaningfully and import meaningfully. Domestic shortage pressure is the trigger for export restriction.

ROLE 04

Surplus potential, reliability constrained

Myanmar · Cambodia · Lao PDR. Real productive capacity limited by logistics, milling, political or landlocked constraints.

ROLE 05

Import-dependent logistics hub

Singapore. Over 90% imported, with regional sourcing, finance, cold chain and trading capability that serve the wider region.

ROLE 06

Import-dependent consumption economy

Brunei · Timor-Leste. High external reliance with limited volume leverage. Timor-Leste is the region's newest and most vulnerable food-security node.

Profile Dimensions

Production strength, import dependence and regional systemic importance by Member State. The three do not move together — which is exactly why a single ranking of "most food secure" would be misleading and is deliberately not shown.

Full Profile Register

StateProduction strengthImport dependenceCommodity concentrationClimate exposureUpstream exposureLogistics exposure
ThailandVery highLowRice, cassava, sugar, poultryMekong, monsoonFertiliser, feedExport ports
Viet NamVery highLow-mediumRice, seafood, coffeeMekong delta, salinityFertiliser, aquafeedExport ports, river
IndonesiaVery highMediumRice, maize, palm oil, fisheriesMonsoon, ENSOFertiliser, feed grainInter-island
MalaysiaMediumMedium-highPalm oil, poultry, eggs, vegetablesMonsoon, floodingFertiliser, feed — highLand corridors, ports
PhilippinesHighMedium-highRice deficit; bananas, coconut, fisheriesTyphoon — very highFeed grain, fertiliserArchipelagic
MyanmarHighLowRice, pulses, maizeMonsoon, cycloneFertiliser, fuelPolitical, logistics
CambodiaMedium-highMediumRice, cassava, maizeMekongFertiliser, milling capacityLand routes to ports
Lao PDRMediumMediumRice, maize, livestockMekongFertiliser, fuelLandlocked
SingaporeVery lowOver 90%All categories importedLow directFeed for local farmsLand corridor, port, air
BruneiLowHighMost categories importedLow directFeed, fuelMaritime
Timor-LesteLowHighStaples importedDrought, rainfall variabilityFertiliser, seed, fuelMaritime, thin integration
Assessments are the platform's characterisation for prototype purposes and are not ASEAN or national government classifications.
DEPENDENCY › PRODUCTION & EXPORTABLE SURPLUS

Production vs Exportable Surplus

The single most important distinction in the ASEAN layer. A country can be an enormous producer and still have almost nothing available for regional substitution, because domestic requirement, mandatory reserves and existing export commitments consume the output before anyone else can reach it.

The Surplus Equation

START
Production
Plus opening inventories, plus imports.
DEDUCT
Domestic consumption
The requirement that is never negotiable in a crisis, and the first thing a government protects.
DEDUCT
Mandatory reserves
National buffer stock requirements that cannot legally or politically be released into export.
DEDUCT
Existing export commitments
Contracts already sold, often to buyers outside ASEAN.
RESULT
Potential Exportable Surplus
The only quantity that can actually answer another Member State's shortfall. This is what the collective substitution engine works with — not production, and not exports.

Production Against Available Surplus

Indexed production against modelled potential exportable surplus for staple grain by Member State. The gap between the two bars is the resilience illusion: it is the volume that exists in the region and cannot be moved to where it is needed. Prototype modelling.

Collective Position by Commodity

CommodityASEAN collective balanceWhere the surplus sitsMobility of that surplusCollective position
RiceNet surplusThailand, Viet Nam, Cambodia, MyanmarGood on the mainland; policy-contingentSURPLUS BUT CONTESTED
MaizeBroadly balancedIndonesia, largely consumed domesticallyLow — feed demand absorbs itTIGHT
Soybean mealNet deficitLargely external to ASEANExternal dependencyEXTERNALLY DEPENDENT
Palm and cooking oilLarge net surplusIndonesia, MalaysiaHigh, but geographically concentratedCONCENTRATED SURPLUS
PoultryBroadly balancedThailand principallyModerate; feed-constrainedFEED-CONSTRAINED
EggsBroadly balancedMalaysia, Thailand, IndonesiaLow — perishable, short rangeSHORT-RANGE
SeafoodNet surplusIndonesia, Viet Nam, ThailandHighRESILIENT
Vegetables and fruitMixed by lineDistributed; highland vegetables concentratedLow — perishabilityPERISHABILITY-LIMITED
WheatNet deficitNot produced at scale in ASEANFully externalEXTERNALLY DEPENDENT
FertiliserNet deficitLargely externalFully externalEXTERNALLY DEPENDENT
Four lines are externally dependent, and three of those — soybean meal, wheat and fertiliser — sit upstream of ASEAN's own production. Regional self-sufficiency in finished food is therefore conditional on inputs the region does not control.
DEPENDENCY › CONCENTRATION & SPOF

Regional Concentration & Single Points of Failure

Two distinct risks that a trade dashboard collapses into one. Import dependency risk is a state relying on outside supply. Production concentration risk is the region relying on a single geography. ASEAN can be collectively self-sufficient in a commodity and still face systemic vulnerability if most surplus originates from one climatic zone or two neighbouring countries.

PADDY CONCENTRATION ASEANstats
90.9%
Five states, 2022
MAIZE CONCENTRATION ASEANstats
52.4%
Indonesia alone, 2022
CASSAVA CONCENTRATION ASEANstats
77%
Four states, 2022
PALM OIL EXPORTS FAO
≈85%
Top two global exporters, 2024

Two Risks, Plotted Separately

Import dependency risk on one axis, production concentration risk on the other. The upper-right quadrant is where both apply — the region depends on outside supply and its own production is concentrated. Prototype scoring.

Regional Single-Point-of-Failure Register

Single pointTypeWhat depends on itRedundancyRating
Mekong basin hydrologyClimaticRice, cassava and maize production across five mainland statesNone — it is a single river systemCRITICAL
Two-country palm oil production baseProduction concentrationRegional and global cooking oil supplyAlternative oils at costCRITICAL
External soybean meal supplyUpstream importPoultry, eggs, pork and aquaculture across the regionLimited regional alternativeCRITICAL
External fertiliser supplyUpstream importCrop yields region-wide, one to three seasons outLimited regional productionCRITICAL
Malacca and Singapore StraitMaritime corridorMost sea-borne food and input flows into maritime ASEANLonger routings at severe costHIGH
Malaysia–Singapore land crossingsBorder corridorFresh vegetables, fruit, eggs and live animals into SingaporeTwo crossings; sea and air at costHIGH
Two dominant rice exportersSupplier concentrationRice supply for at least four deficit statesCambodia, Myanmar, external originsHIGH
Archipelagic logisticsDomestic distributionInternal food movement in Indonesia and the PhilippinesMultiple routes, capacity-limitedELEVATED
Regional cold chain and powerInfrastructureSeafood, meat, eggs, perishables, frozen reservesDistributed but power-dependentELEVATED
Six of the nine entries are shared by multiple Member States. That is the definition of a regional single point of failure: no country can retire it alone, and no country can be insured against it by national diversification.
DEPENDENCY › CORRELATED EXPOSURE

Correlated Exposure & False Diversification

This engine is the direct regional extension of the correlated country exposure logic already running in the Singapore layer, where a single Malaysian event scores simultaneously against eggs, vegetables, fruit, seafood, poultry and live animals rather than as six independent readings.

Correlation Factors Modelled

CLIMATIC

Shared systems

Mekong hydrology · monsoon systems · ENSO cycles · shared river basins · drought and flood geography.

BIOLOGICAL

Shared disease geography

Transboundary animal disease · plant disease and pests · aquatic disease · vector distribution.

UPSTREAM

Shared inputs

Fertiliser · imported maize · soybean meal · energy · seeds · water · aquafeed.

STRUCTURAL

Shared infrastructure and policy

Shipping lanes · ports · border crossings · cold chain · export policy reflex · regulatory authority.

False Diversification Detector

Apparent diversification against effective diversification for each substitution pair, after correlated factors are removed. Where the two bars diverge sharply, the diversification is nominal. Prototype scoring.

Worked Cases

Apparent diversificationShared factorEffective diversificationVerdict
Poultry from producer A switched to producer BBoth depend heavily on imported soybean meal and maizeLow against a feed shockFALSE DIVERSIFICATION
Rice from Thailand switched to Viet Nam or CambodiaAll three draw on the Mekong basin and the same monsoon systemLow against a regional climate eventFALSE DIVERSIFICATION
Vegetables from one regional highland switched to anotherBoth depend on imported fertiliser and the same rainfall regimePartialPARTIAL
Eggs from a regional source switched to local productionBoth depend on the same imported feed grainPartial — protects against border risk, not feed riskPARTIAL
Seafood from Indonesia switched to Viet NamDifferent basins, different species mix, different disease profileHighGENUINE
Frozen protein from a regional source switched to a distant originDifferent climate, different disease geography; shares only shippingHighGENUINE
Rice from the Mekong switched to South Asian originsDifferent monsoon system; shares export-policy behaviourGood against climate, weak against policyPARTIAL
Note the last row. Diversification can be genuine against one factor and false against another, which is why the correlation factor is scored per shock class rather than as a single number per supplier pair.
DEPENDENCY › UPSTREAM DEPENDENCY GRAPH

ASEAN Upstream Dependency Graph

A food security platform that monitors only finished food products will detect many crises too late. Growing feed demand is a major driver of international agricultural trade, and the inputs behind ASEAN's own production are where a distant shock first enters the regional food system.

Chain 01 — Fertiliser to crops

External fertiliser supply
Largely imported into ASEAN; price and availability set outside the region.
Crop yield and planted area
Production cost rises, application rates fall, area contracts.
Rice · vegetables · fruit · maize
Volume and quality fall one to three growing seasons later.

Chain 02 — Feed to protein

Maize + soybean meal
Substantially external to ASEAN; global price transmission is fast.
Compound animal feed
Producer margins compress across several Member States simultaneously.
Poultry · eggs · pork · aquaculture
Four protein networks degrade together, then domestic price pressure drives trade action.

Chain 03 — Energy to everything

Energy
Fuel and power pricing across the region.
Fertiliser · processing · cold chain
Energy is an input to fertiliser manufacture, to processing, and to every refrigerated node.
Production cost and perishable integrity
Raises cost at every stage and threatens the cold chain that perishables and frozen reserves depend on.

Chain 04 — Shipping to trade

Shipping capacity and fuel
Freight rates, vessel availability, chokepoint state.
Input import and food export
Both directions affected: inputs cannot arrive and surplus cannot move.
Regional substitution capability
A surplus that cannot be shipped is not substitution capacity.

Upstream Transmission Timing

Time from an upstream shock to a visible effect on finished food supply, by input class. Fertiliser carries the longest lag and therefore the longest warning window; shipping carries the shortest. Monitoring only the right-hand side of this chart guarantees late detection.

Upstream Watch Register

InputFeedsASEAN positionLag to food effectWatch state
FertiliserAll crop productionNet importer1–3 seasonsRED
Soybean mealPoultry, eggs, pork, aquacultureNet importer1–2 quartersRED
MaizeCompound feedBroadly balanced, domestically absorbed1–2 quartersAMBER
Energy and fuelFertiliser, processing, cold chain, shippingMixed; several net importersWeeks to quartersAMBER
Seeds and planting materialCrop productionPartly imported1 seasonAMBER
Water and river flowRice, agriculture, aquacultureShared and transboundaryWithin seasonRED
Shipping and freightInput import, food exportRegionally strong, externally exposedWeeksAMBER
Cold storage capacitySeafood, meat, perishables, reservesDistributed, power-dependentImmediateAMBER
DEPENDENCY › CORRIDORS & CHOKEPOINTS

Regional Corridors & Chokepoints

Food security is a network flow problem, not a set of trade statistics. Perishables in particular must move through a defined chain — production node, aggregation node, border, port, shipping or road corridor, cold store, importer, consumer market — and each link fails differently.

Corridor Register

CorridorTypeCarriesDominant riskRedundancyState
Malaysia–Singapore land crossingsBorderFresh vegetables, fruit, eggs, live animals, poultryBorder clearance, policy, congestionTwo crossingsRED
Malacca and Singapore StraitMaritimeMost sea-borne food and inputs into maritime ASEANCongestion, incident, securityEffectively none at scaleAMBER
Mekong economic corridorsRoad and riverRice, cassava, maize, livestock across the mainlandWater level, road capacity, border frictionPartial road alternativesAMBER
Thailand–Malaysia land corridorBorderRice, produce, processed food southboundBorder processing, security, seasonal floodingSea alternative at costGREEN
Viet Nam and Thailand export portsPortRegional and global rice and seafood exportsCongestion, capacity, typhoonMultiple portsGREEN
Indonesia inter-island logisticsDomestic maritimeInternal distribution of a huge production baseVessel capacity, weather, costMany routes, capacity-limitedAMBER
Philippines archipelagic logisticsDomestic maritimeInternal staple distribution across the islandsTyphoon, port damage, costWeather-contingentRED
Brunei maritime dependenciesMaritimeNearly all imported requirementSmall volumes, limited direct callsTranshipment via neighboursAMBER
Timor-Leste maritime supplyMaritimeStaple imports for the newest Member StateThin service frequency, port capacityVery limitedRED
Regional cold chain nodesInfrastructureSeafood, meat, eggs, perishables, frozen reserveCapacity, power continuity, temperature integrityDistributedAMBER

Five Distinct Failure Types

Production risk, route risk, border risk, port risk and capacity risk scored separately by corridor. They are deliberately not aggregated: the response to a border problem is diplomatic, to a port problem is logistical, and to a capacity problem is capital. Conflating them produces the wrong playbook.
RISK & DECISION › ASEAN RISK MATRIX

ASEAN Food Security Risk Matrix

Country against commodity, scored by the patented kernel. Every value shown here conceptually originates from SG 10202109991Y — no AI model, LLM or agent may originate, change or override an authoritative risk state. The matrix distinguishes an isolated national shortage from a disruption to a systemically important supplier.

MEMBER STATE × COMMODITY — KERNEL RISK STATE ILLUSTRATIVE PROTOTYPE VALUES
KERNEL STATE LOWINTERMEDIATEHIGH

Three Layers of the Same Question

LAYER 01

Country × commodity

Is this state at risk on this commodity? The national view, shown above. Answers whether a shortage is likely.

LAYER 02

Country → country dependency

Is this state's risk transmissible to another? Shown in the trade matrix. Answers whether a shortage spreads.

LAYER 03

Regional systemic importance

Would this state's disruption destabilise the region? Answers whether it is a national event or an ASEAN event — and therefore who needs to be in the room.

Systemic Importance Weighting

Each Member State's disruption impact, weighted by how much other states depend on it. A high national risk in a low-systemic-importance state is a humanitarian and national issue; a moderate risk in a high-systemic-importance state may be the larger regional problem. Both matter, and they call for different instruments.
RISK & DECISION › INTERDEPENDENCY INDEX

ASEAN Food Interdependency Index

A composite score computed for every exporter, importer and commodity triple. It exists so that a decision maker can rank dependencies rather than merely list them — and so that the effect of a policy intervention can be measured as a change in the index rather than as an assertion.

AFII — CONCEPTUAL FORMULATION PROTOTYPE SCORING MODEL, TO BE CALIBRATED WITH ACTUAL DATA
AFII  =  Import dependence
      × Supplier concentration
      × ASEAN systemic importance
      × Correlated risk factor
      × Recovery time
      × Logistics fragility
      − Substitution capacity
      − Inventories and reserves
      − Domestic production resilience
This is a prototype conceptual scoring model. It is presented as a structure to be calibrated against an assembled eleven-country by eleven-country by commodity dataset, not as a validated index. The multiplicative terms are the ones that compound a problem; the subtractive terms are the ones a policy intervention can move.

Highest AFII Dependencies

Ranked AFII scores for the highest-scoring exporter–importer–commodity triples. Prototype values. The ranking is the point: it converts a wall of trade statistics into a queue of things worth intervening on.

Component Contribution

ComponentWhat it capturesData source in deploymentMoveable by policy?
Import dependenceShare of national requirement met by importsProduction, consumption, trade dataSlowly — via domestic production
Supplier concentrationTop-1 and top-3 share; HHI across suppliersBilateral trade cubeYes — accreditation and sourcing
Systemic importanceHow many other states depend on the same exporterNetwork centrality of the exporterIndirectly
Correlated riskShared climate, basin, disease, feed, fertiliser, route, authorityCorrelation engineYes — by choosing genuinely uncorrelated alternates
Recovery timeHow long the commodity takes to come back after a shockBiological and agronomic parametersNo
Logistics fragilityCorridor, port, border and cold chain exposureCorridor register, AIS, port dataYes — infrastructure and routing
Substitution capacityExportable surplus reachable in time and eligibleSurplus model plus substitution engineYes — the primary lever
Inventories and reservesNational stocks plus regional reserve mechanismsNational reserve data, APTERR positionYes — reserve policy
Production resilienceDomestic capacity to absorb a shortfallProduction and yield dataYes — long horizon
Five of the nine components are directly moveable by policy. That is the practical value of decomposing the index: it tells a minister not only how bad a dependency is, but which of the levers available to them would actually reduce it.
RISK & DECISION › COLLECTIVE SUBSTITUTION ENGINE

Collective Substitution Engine

Do not optimise Member States independently. If one major rice supplier falls, the Philippines, Malaysia, Singapore and Brunei may each be advised to buy from the same alternative. Every recommendation is individually rational. Collectively they may be impossible. The engine exists to detect that before governments and commercial buyers compete for an insufficient substitute supply.

The Headroom Calculation

START
Available regional exportable surplus
From the surplus model — not production, and not historical exports.
DEDUCT
Existing commitments
Contracts already sold, including to buyers outside ASEAN.
DEDUCT
Domestic reserve requirements
The buffer each exporting state must legally or politically retain.
DEDUCT
Simultaneous substitution demand
The sum of what every affected Member State would seek from this source at the same time. The term that national models omit.
RESULT
ASEAN Regional Substitution Headroom
The volume that can actually be redirected. If it is negative, no combination of independently rational national decisions can resolve the shortfall, and a coordinated instrument is required.

Worked Case — Loss of a Major Rice Supplier

Regional headroom as each affected Member State enters the market for the same alternative. Headroom is adequate for the first claimant and negative by the fourth. Prototype modelling. The chart is the argument for coordination: the fourth state is not badly managed, it is simply late.

Nine-Step Determination

StepQuestionWhy it can fail
1Volume affectedMeasured against national requirement, not against the exporter's output
2Which importing Member States are exposedExposure is often broader than the obvious bilateral relationship
3Which ASEAN states hold available exportable surplusProduction is not surplus; surplus is what remains after domestic needs and commitments
4Non-ASEAN alternative suppliersExternal origins may face their own restriction or price regime
5Logistics feasibilityCorridor, freight capacity, cold chain and transit time inside the window
6Food safety and SPS feasibilityAn origin that is available but not eligible is not an option
7Time to substitutionAccreditation, contracting, shipping and clearance all consume the window
8Price implicationsSimultaneous demand moves the price against every claimant at once
9Impact of all affected states substituting simultaneouslyThe step national models omit, and the reason individually rational advice becomes collectively impossible

Ranked Collective Response Options

RankOptionInstrumentSpeedCollective harmAssessment
1Sequenced allocation of available surplusCoordination between exporting and importing statesDaysLowestPREFERRED
2Regional reserve release to the most exposed statesEmergency reserve mechanismsDays to weeksLowVIABLE
3External sourcing for the states with the most buying powerCommercial procurement, diplomatic facilitationWeeksModerate — raises world priceCONDITIONAL
4Temporary regulatory facilitationAccelerated clearance, recognition of equivalent certificationWeeksModerateCONDITIONAL
5Logistics prioritisationCorridor, berth and cold chain allocationWeeksModerateCONDITIONAL
6Production surgePlanting incentives, input allocationSeasonsLow but slowLONG HORIZON
7Uncoordinated national procurementEach state acting aloneImmediateHighest — price spiral, late claimants unservedDEFAULT IF NOTHING IS DONE
The last row is not a recommendation. It is what happens in the absence of a shared picture, and it is the outcome the engine exists to make visible before it occurs.
RISK & DECISION › DISRUPTION SIMULATOR

ASEAN Disruption Simulator

Six prototype scenarios, each traced from the directly affected states through second-order propagation to the recommended collective response. The purpose is not prediction. It is to establish, before an event, which instruments would matter and which states would need to be in the room.

Scenario Propagation

Number of Member States directly affected, second-order affected and the modelled regional headroom remaining, by scenario. Prototype modelling.
SCENARIO A

Major rice production shock in a leading exporter

Directly affected: the exporting state.
Second order: Philippines, Malaysia, Singapore, Brunei, Timor-Leste — all seeking the same alternates.

Consequences: demand surge onto remaining mainland exporters, prices rise, spare capacity in smaller exporters is exhausted, freight demand rises, reserve activation probability rises.

The canonical simultaneous substitution case. Headroom turns negative before the fourth claimant is served.
SCENARIO B

A Member State introduces temporary food export restrictions

Directly affected: its immediate import-dependent neighbours.
Propagates through: eggs, vegetables, fruit, poultry and seafood simultaneously, because a single authority governs all of them.

Response: substitution calculated across the remaining regional producers and external suppliers, with diplomatic engagement ranked ahead of commercial substitution because the constraint is political.

Directly mirrors the documented 2022 poultry episode. The correlated multi-commodity effect is the point.
SCENARIO C

Severe Mekong drought

Simultaneously affected: Thailand, Viet Nam, Cambodia and Lao PDR, across rice, cassava and maize.

Why it is different: country diversification offers no protection, because the diversified suppliers share the basin. Regional climate concentration is the exposure.

Demonstrates why the correlation engine exists. This scenario invalidates most national substitution plans at once.
SCENARIO D

Global soybean meal and feed shock

Propagates into: poultry, eggs, pork and aquaculture across Thailand, Malaysia, Indonesia, Viet Nam and the Philippines simultaneously.

Signature: no food commodity fails first. Producer margins compress, then domestic prices rise, then export restrictions follow months later.

The clearest demonstration of the upstream concept: the longest warning window in the set, and the one most often missed.
SCENARIO E

Major transboundary animal disease in regional poultry

Directly affected: producing states inside the outbreak zone.
Second order: every importing state relying on them, plus egg supply through the same flocks.

Instruments: regionalisation to preserve unaffected zones, accelerated accreditation of alternates, form substitution to frozen.

Recovery is measured in flock rebuild time, not shipping time — which is why early detection is worth disproportionately more here.
SCENARIO F

Malacca and Singapore maritime disruption

Directly affected: maritime ASEAN food and input flows.
Second order: mainland states dependent on those ports for inputs and for export routing.

Signature: both directions fail at once — inputs cannot arrive and surplus cannot move, so regional substitution capacity itself is degraded.

The scenario in which a healthy surplus somewhere in the region is irrelevant, because it cannot reach the state that needs it.

What Each Scenario Run Produces

Directly affected statesSecond-order affected statesCommodities in scope Volume affectedRisk propagation pathPrice pressure ASEAN substitution headroomExternal sourcing requirementRegional reserve implications Recommended collective responseInstruments and their ownersKernel risk state per state and commodity
Every scenario output passes through the same chain as a live event: sense, verify, kernel score, cross-border inference, collective decision, agentic staging, human authorisation. A simulation and a real event differ only in whether the trigger is hypothetical.
RESPONSE › ASEAN EARLY WARNING

ASEAN Early Warning

The same sensing architecture that runs the national layer, scaled to eleven Member States and their external suppliers. The chain is unchanged and non-negotiable: signal, verification, patented kernel, cross-border inference, ASEAN impact, response options, human decision.

STAGE 01
Signal
Weather · ENSO · Mekong hydrology · crop conditions · harvest forecasts · animal disease · plant disease · commodity prices · fertiliser · feed · energy · freight · ports · AIS shipping · border conditions · national policy · export restrictions · food inflation · social and news signals.
STAGE 02
Verification
Adversarial multi-agent assert, falsify and source-reliability scoring, in the languages of every Member State. In a regional setting a false positive can itself move a market, so verification precedes scoring absolutely.
STAGE 03 · SG 10202109991Y
Patented risk determination kernel
Trigger event validated, compared against geospatial and weather data in real time, risk value computed with positive and negative components, impact percentile tested against threshold, classified high, intermediate or low, then supply-side or demand-side. The only authoritative source of a risk value in the platform.
STAGE 04
Cross-border inference
Propagates the scored event across the ASEAN ontology to every affected importing state, commodity, corridor, reserve line and downstream market — including second-order effects across time.
STAGE 05
ASEAN impact and collective options
Headroom computed, simultaneous substitution risk tested, ranked collective response options generated with instruments and owners identified.
STAGE 06
Human decision
Member States and ASEAN bodies decide. The platform stages; it does not act. National sovereignty and institutional authority are preserved absolutely.

Regional Signal Feed

DetectedSignalClassStates implicatedConfidenceKernel state
T−0hMekong basin flow below seasonal normal across multiple reaches
Correlated across four mainland states rather than one
HydrologyTH · VN · KH · LA0.83HIGH
T−9hFertiliser price and freight dislocation persisting
Transmission to yields lags one to three seasons
Upstream inputRegion-wide0.88INTERMEDIATE
T−1dSoybean meal forward curve steepening
Four protein networks share this input
Feed marketTH · MY · ID · VN · PH0.91HIGH
T−2dExport restriction language in a Member State ministry statement
Matches the documented pre-restriction pattern
Trade policy1 exporting state0.62INTERMEDIATE
T−3dTransboundary poultry disease notifications clustering
Zone extent still being resolved
Animal health2 producing states0.74INTERMEDIATE
T−5dTyphoon track projection over an archipelagic staple region
Port damage and internal distribution risk
ClimatePH0.79HIGH
T−9dMaritime service frequency thinning on a small-state supply route
Affects the least integrated Member State disproportionately
LogisticsTL0.66INTERMEDIATE
Illustrative prototype feed. Confidence is the post-verification posterior. Every kernel state shown is produced by the patented kernel, never by the sensing or language layer.
RESPONSE › REGIONAL RESERVES & APTERR

Regional Reserves & APTERR

APTERR, AFSRB and AFSIS are existing ASEAN mechanisms with their own governance and authority. This platform neither owns, controls nor speaks for any of them. It models them as decision instruments so that an officer can see what each could deliver, how fast, and to whom — nothing more.

The ASEAN Plus Three Emergency Rice Reserve is a regional cooperation established under the APTERR Agreement signed by the Ministers of Agriculture and Forestry of ASEAN Plus Three on 7 October 2011 in Jakarta, with the Secretariat established in Bangkok in 2013. It is the region's most developed operational food security instrument, and the platform models it as an actual decision instrument rather than as another dataset.

EARMARKED STOCKS APTERR
787,000t
Total earmarked emergency rice reserve
ASEAN CONTRIBUTION APTERR
87,000t
From ASEAN Member States
PLUS THREE CONTRIBUTION APTERR
700,000t
China 300,000 · Japan 250,000 · ROK 150,000
PRE-POSITIONED APTERR
7,522t
In Lao PDR, Myanmar and the Philippines

Reserve Structure

Earmarked contributions by member. The structural observation for a decision maker: the great majority of earmarked volume comes from the Plus Three partners, so ASEAN's principal rice emergency instrument is a partnership rather than a purely intra-ASEAN capability.

The Three Tiers as Decision Instruments

TierMechanismSpeedBest suited toLimitation
Tier 1Forward trading — rice delivered from the supplying country to the recipient during an emergency, on agreed commercial termsWeeksAnticipated shortfalls with a known requirementRequires the recipient to transact; not a grant
Tier 2Earmarked stock release arrangementsWeeksAugmenting national food security reserves of an affected stateVolume constrained by earmarked position
Tier 3Grant and pre-positioned programme, distributed by the national food security agency of the affected memberDaysAcute disaster response — flood, typhoon, earthquakeModest volumes relative to national consumption
Stocks released by APTERR are distributed by the national food security agency of the affected member country. Tier 1 and Tier 2 releases would typically augment the national food security reserves of that state rather than reach consumers directly.

Reserve Modelling in the Platform

Modelled quantityWhat the officer seesWhy it matters
Reserve availabilityEarmarked and pre-positioned position against the modelled scenarioDistinguishes a pledge from rice that is physically nearby
Eligible emergency requirementWhat the scenario would qualify for under existing mechanismsPrevents planning around volumes that would not be released
Country requirementThe affected state's shortfall in tonnes and in days of consumptionConverts a reserve figure into an operationally meaningful duration
Release timingTier-dependent lead time from request to distributionDetermines whether the reserve is a bridge or arrives after the gap
LogisticsCorridor, port and internal distribution feasibilityA reserve that cannot be moved is not a reserve
Days of consumption supportedThe headline number for a decision makerThe only figure that answers "does this buy us enough time?"
Since its establishment the mechanism has supported disaster response across the region, including flood, typhoon and earthquake responses in Lao PDR, Myanmar and the Philippines. Its design purpose is emergency response, and the platform models it accordingly — as a bridge instrument, not as a substitute for trade.
RESPONSE › COLLECTIVE RESPONSE PLAYBOOKS

Collective Response Playbooks

Ten regional playbooks. Each is triggered by a kernel-determined risk state, supported by cross-border inference, staged by agents, and authorised by the appropriate human decision makers in ASEAN bodies and Member State governments. The platform prepares; it never issues.

PB-A1 · ARMED

Regional Rice Supply Shock

Trigger: kernel scores HIGH on rice for a systemically important exporter.

Stages: headroom computation, sequenced allocation, reserve eligibility assessment, external sourcing options.

Authority: Member States and relevant ASEAN bodies.

PB-A2 · ARMED

Export Restriction Coordination

Trigger: pre-restriction policy signalling detected in a Member State.

Stages: affected-state mapping, diplomatic option set, substitution pre-positioning.

Authority: national governments; regional consultation.

PB-A3 · ARMED

Transboundary Animal Disease

Trigger: outbreak scored HIGH with cross-border propagation projected.

Stages: zone resolution, regionalisation evidence, alternate accreditation ranking, form substitution.

Authority: national veterinary authorities.

PB-A4 · ARMED

Mekong Climate Shock

Trigger: basin hydrology below threshold across multiple reaches.

Stages: correlated exposure assessment, multi-state impact projection, external sourcing requirement, reserve pre-assessment.

Authority: riparian states and regional bodies.

PB-A5 · ARMED

Regional Feed & Fertiliser Shock

Trigger: upstream input index crosses the historical intervention band.

Stages: multi-network propagation model, producer margin assessment, restriction probability forecast.

Authority: national agriculture and trade ministries.

PB-A6 · ARMED

Critical Logistics Corridor Failure

Trigger: a corridor in the register turns red.

Stages: reroute modelling, capacity allocation, cold chain integrity assessment, both-direction impact.

Authority: transport ministries and port authorities.

PB-A7 · STANDBY

Simultaneous Substitution Demand

Trigger: regional headroom projected to turn negative under current national intentions.

Stages: claimant sequencing, allocation proposal, price impact model, coordination convening.

Authority: regional coordination; national consent.

PB-A8 · STANDBY

Emergency Reserve Assessment

Trigger: an affected state's coverage falls below the critical band with no substitution path.

Stages: eligibility assessment, tier selection modelling, logistics feasibility, days-of-consumption calculation.

Authority: APTERR governance and the affected Member State.

PB-A9 · ARMED

External Supplier Diversification

Trigger: regional surplus insufficient for the modelled requirement.

Stages: external origin ranking, SPS and eligibility screening, lead time and price modelling.

Authority: national procurement and trade authorities.

PB-A10 · ARMED

Regional Food Price Shock

Trigger: food inflation and price transmission cross threshold across multiple states.

Stages: transmission mapping, vulnerable population exposure, demand-side management options, public communication framing.

Authority: national governments.

GOVERNING RULE

Authority boundary

No playbook executes. Every one is staged for human decision makers who hold the relevant institutional authority. The platform computes, ranks and prepares. Member States and ASEAN bodies decide and act.

Playbook Trigger Map

Detection lead available against coordination time required, by playbook. Playbooks above the diagonal can be convened before the disruption binds; those below can only be run reactively. Moving a playbook above the line is a sensing investment, not a diplomatic one.
INSTITUTIONS › INSTITUTIONAL LAYER

ASEAN Institutional Layer

ASEAN already has the institutional architecture for regional food security. What it does not yet have is a continuous, scored, cross-border analytical layer connecting those mechanisms to each other in real time. The platform positions itself as decision-intelligence connective tissue between existing bodies — never as a replacement for their authority.

Existing Mechanisms and the Platform's Relationship to Each

InstitutionEstablished rolePlatform relationshipWhat the platform contributes
AMAFASEAN Ministers on Agriculture and Forestry — policy and governanceSERVESA scored regional picture to inform ministerial decisions
SOM-AMAFSenior officials preparing and implementing AMAF directionSERVESContinuous monitoring between meetings rather than periodic reporting
AFSISASEAN Food Security Information System — systematic collection, analysis and dissemination of food security informationCOMPLEMENTSAn analytical and decision layer on top of information AFSIS already assembles
AFSRBASEAN Food Security Reserve Board — regional food security monitoringCOMPLEMENTSScenario modelling of reserve adequacy against scored risk
APTERRASEAN Plus Three Emergency Rice Reserve — emergency rice interventionMODELS AS INSTRUMENTEligibility, timing, logistics and days-of-consumption modelling; no ownership or control
ASEAN SecretariatRegional coordinationSERVESA shared operating picture across Member States
National food and agriculture ministriesNational executionSERVESNational view within regional context; sovereignty preserved
Trade ministries and customsTrade policy and clearanceSERVESRestriction detection, corridor state, clearance friction
Transport ministries and port authoritiesCorridor and port operationSERVESCorridor risk and reroute modelling
National statistics agenciesProduction, consumption and trade dataDEPENDS ONConsumes their data; never becomes a competing record of truth
Private producers, importers and exportersActual sourcing, production, shipping and substitutionSERVES AND LEARNS FROMSubstitution actually happens here; the enterprise tier closes the loop
Shipping, logistics and trade associationsMovement and coordinationSERVESCorridor intelligence and coordinated switching across members
The distinction between SERVES, COMPLEMENTS and MODELS AS INSTRUMENT is deliberate and load-bearing. It is the answer to the first question any ASEAN body will ask, which is whether this proposes to take authority away from them. It does not.

Where the Platform Sits

EXISTING
Information mechanisms
AFSIS, national statistics agencies, national ministries — the sources.
NEW LAYER
ASEAN Collective Food Security AIC³
Sense · verify · map dependency · score risk through the patented kernel · infer cross-border consequences · simulate · identify substitution · generate collective response options · coordinate.
EXISTING
Decision and intervention mechanisms
AMAF, SOM-AMAF, AFSRB, APTERR, national governments — the deciders and the instruments.
The platform occupies the gap between information and decision. That gap is where regional food security is currently lost: the data exists, the instruments exist, and the continuous analytical connection between them does not.
INSTITUTIONS › ASEAN 2045 ALIGNMENT

Policy Alignment

ASEAN has already adopted the objectives this platform implements. The proposition is not a new policy direction but a decision-intelligence implementation layer for commitments the region has formally made.

FRAMEWORK 01

ASEAN Community Vision 2045

The long-horizon vision within which the sectoral and economic plans sit. Timor-Leste committed to the ASEAN Charter and Vision 2045 on admission.

FRAMEWORK 02

Food, Agriculture and Forestry Sectoral Plan 2026–2030

Calls for a sustainable, inclusive, resilient and secure regional food system, aligned with Vision 2045.

FRAMEWORK 03

AEC Strategic Plan 2026–2030

Calls explicitly for integrated food security information systems capable of forecasting, planning and monitoring supplies and utilisation of basic food commodities; stronger emergency reserve mechanisms; accelerated digital transformation; and public-private partnerships.

Objective-by-Objective Alignment

Adopted ASEAN objectivePlatform capabilityView
Strengthen food security and nutritionContinuous scored risk across eleven states and every major commodity01 · 11
Strengthen integrated food security information systemsThe bilateral trade cube, national ontology and regional data spine03 · 20
Forecast, plan and monitor supplies and utilisation of basic food commoditiesProduction, consumption, exportable surplus and headroom modelling06 · 13
Strengthen resilient supply chainsCorrelated exposure, upstream dependency and corridor risk engines08 · 09 · 10
Enhance trade and market connectivityInterdependency map and intra-ASEAN trade matrix02 · 03
Strengthen emergency reserve arrangementsReserve modelling against scored scenarios, with tier and timing analysis16
Accelerate digital transformationThe platform itself — an operating capability rather than a reportAll
Promote public-private partnershipsThe enterprise tiers, where firms receive intelligence and return demand signalsIndustry portal
Reduce dependence on concentrated partners in key sectorsConcentration, SPOF and AFII engines; ranked diversification candidates07 · 12
Nine adopted objectives, nine implemented capabilities. This alignment is the strongest argument for the regional layer: the strategic case has already been made and accepted by ASEAN. What is missing is the operational instrument.

Positioning

Singapore's roleA logical technology, analytics, logistics, finance and decision-intelligence hub from which a regional food security capability can be demonstrated — and the national reference implementation that proves the method works.
Member State authorityPreserved entirely. ASEAN institutions and Member States retain their respective decision-making authority. The platform implies no centralised control.
Intellectual propertyThe kernel rests on granted Singapore patent SG 10202109991Y, whose priority application was titled Food Security Exchange System and filed with IPOS in September 2020. Regional deployment would proceed by licence, with all national and regional data remaining the property of the contributing authority.
What would make it realAn assembled eleven-country by eleven-country by commodity dataset from ASEANstats, AFSIS, FAOSTAT, UN Comtrade and national customs sources, with importer dependence, exportable surplus, concentration and network centrality calculated from actual values. That step converts a persuasive demonstrator into a genuine ASEAN Food Interdependency Model.
INSTITUTIONS › ASEAN DATA SPINE

ASEAN Food Data Spine

The module must not be built around a hard-coded trade table. It is built around a regularly refreshed bilateral trade cube, so that every illustrative array in this prototype can be replaced by an API or database feed without redesigning the analytics above it.

The Trade Cube

DIMENSION 01

Exporter

Eleven ASEAN Member States plus external suppliers — India, Pakistan, Australia, Brazil, the United States, China and others.

DIMENSION 02

Importer

Eleven ASEAN Member States.

DIMENSION 03

Commodity

HS6 initially, mapped upward into operational food security groups.

DIMENSION 04

Time

Annual, then quarterly, then monthly where available.

DIMENSION 05

Measure

Value · tonnage · share of importer's imports · share of exporter's exports · unit value · production · consumption · stocks · exportable surplus.

Operational Commodity Groups

RiceWheatMaizeSoybeans and feed PoultryEggsPorkBeef SeafoodVegetablesFruitSugar Cooking oilDairyProcessed staple foods FertiliserAnimal feed
Seventeen operational groups, of which two — fertiliser and animal feed — are not food at all. They are in the cube because they are where a regional food crisis begins, and excluding them would guarantee late detection.

Source Register

SourceProvidesRefreshRole
ASEANstatsRegional trade and production aggregatesPeriodicRegional baseline
AFSISASEAN food security information, systematically collected and disseminatedPeriodicRegional food security spine
FAOSTAT detailed trade matrixBilateral commodity tradeAnnualBilateral cube backbone
UN ComtradeHS-level bilateral tradeMonthly to annualGranularity and recency
National customs and statistics agenciesAuthoritative national trade and productionMonthlyGround truth
National food reserve and production dataStocks, reserves, harvestVariesSurplus and headroom calculation
WITS and BACIReconciled historical tradeAnnualHistorical analysis and backtesting
Weather and climate systemsENSO, hydrology, rainfall, crop conditionDailyClimate correlation engine
WOAH animal health dataOutbreak notifications and zonesContinuousDisease propagation model
IPPC plant health dataPhytosanitary notificationsContinuousPlant health watch
AIS and shipping dataVessel movement, chokepoint transitsContinuousCorridor and route model
Commodity, feed and fertiliser marketsPrice and forward curvesDailyUpstream early warning
Private logistics and industry dataFreight capacity, cold chain, actual sourcing outcomesContinuousFeasibility and feedback loop

Regional Ontology

CountryCommodityProductionConsumption ImportExportSupplierImporter FarmProcessorPortBorder RouteStockpileReserveClimate region River basinDiseaseFeedFertiliser EnergyPolicyTrade agreementResponse instrument
Twenty-four entity types. The ontology is what allows a single scored event to propagate correctly — a river basin connects to five countries, a feed input connects to four protein networks, and a policy decision connects to every commodity under one authority. Without the graph, the correlation engine has nothing to reason over.

Data Integrity Rules

Never label prototype data as officialNo illustrative figure is presented as an ASEAN, ASEANstats, AFSIS, APTERR, FAO or Member State statistic. Published figures carry a source tag; everything else carries an illustrative tag.
Replaceable arraysEvery illustrative array is structured so it can be replaced by a live feed without changing the analytics above it.
Provenance on every risk valueEach kernel output carries its evidence chain, model version and the data vintage behind it.
National data sovereigntyContributing authorities retain ownership of their data. The platform reads from systems of record and never becomes a competing one.
Aggregation thresholdsFirm-level and commercially sensitive positions are never exposed across borders; only aggregates above a minimum count are published.
The next research stepAssemble the actual eleven by eleven by commodity dataset and calculate importer dependence, exportable surplus, concentration and network centrality from real values. That is what turns this demonstrator into an analytical model.
AFSIM COMMAND › EXECUTIVE DASHBOARD

AFSIM — ASEAN Food Security Index & Monitor

Regional food security intelligence, early warning, scenario analysis and decision support. AFSIM is the measurement and monitoring layer that sits over the ASEAN collective views: it turns the region's fragmented food, agricultural, trade, climate, market, disease and logistics information into one common picture of risk, vulnerability and resilience — and connects that picture to warning, simulation and coordinated response.

Independent prototype — read the badges. AFSIM is not an ASEAN system and carries no ASEAN, ASEANstats, AFSIS, APTERR, FAO, WOAH or Member State endorsement. It is an analytical and decision-intelligence layer intended to complement, not replace, existing national and ASEAN mechanisms, and it implies no centralised control over Member States. Every number below carries a confidence badge. Nothing tagged MODELLED or ILLUSTRATIVE is an ASEAN statistic, and no figure on this platform is presented as official.

AFSIM COMPOSITE INDEX
OF 100 · HIGHER IS MORE SECURE

The state of ASEAN food security, September 2026

Computing from the data register…

MEMBER STATES AT ELEVATED RISK
Country AFSIM below the regional score
ACTIVE WARNINGS
Above threshold in the early warning engine
REGISTER COVERAGE
Indicators scored from data, not assumed

ASEAN Food Security Scorecard

DimensionASEANStrongest signalWeakest signalCoverageTrend
Click any dimension to open the indicators behind it. Each dimension score is the mean of its scored indicators; indicators marked DATA GAP are excluded from the mean and reduce the coverage figure rather than being silently filled in. The composite is the equally weighted mean of the five dimension scores. Weights, normalisation bands and the full derivation are published in AFSIM Methodology.

What Moved the Index

DOMINANT DRIVER — DOWNWARD

A strengthening El Niño over a concentrated production base

NOAA's Climate Prediction Center holds an El Niño Advisory with a greater than 90 per cent chance of a very strong event through the northern autumn and winter of 2026-27. PUBLIC The ASEAN Specialised Meteorological Centre put the Niño3.4 index at 1.3 °C in July 2026 and expects further strengthening, with a positive Indian Ocean Dipole developing over the same window — a configuration that historically compounds the drying effect across the region. PUBLIC

The exposure is structural rather than incidental: five Member States account for 90.9 per cent of ASEAN paddy production. DERIVED A regional climate signal is therefore not eleven separate national risks — it is one risk with eleven balance sheets attached.

DOMINANT DRIVER — UPSTREAM

Fertiliser, one growing season behind the headline

The closure of the Strait of Hormuz on 28 February 2026 placed roughly a third of global seaborne fertiliser trade at risk, blocking about 21 million tonnes of annual urea and 4 million tonnes of DAP export capacity. PUBLIC Urea roughly doubled to above US$850 per tonne in April before easing to US$453 by June; DAP rose from about US$580 to about US$770. PUBLIC

Fertiliser is in the commodity universe precisely because it is not food. Input cost and availability reach yields one to three growing seasons out, which means the upstream dimension is the earliest reliable warning AFSIM produces — and it is currently the second weakest.

So What — The Decision Chain

ObservationConsequence AFSIM computesWhat ASEAN could consider
Regional rice output could fall 2–8%
MODELLED
The 5% midpoint is roughly equivalent to removing Thailand's entire annual export programme of 7 to 8 million tonnes from the market. Importing states would seek substitutes simultaneously.Quantify substitution headroom before the October ASMC outlook, not after
APTERR earmarked reserve is 787,000 t
VERIFIED
Against combined Southeast and East Asian consumption of about 542,000 tonnes a day, that is 1.45 days of coverage. The mechanism is a signalling and humanitarian instrument, not a supply buffer.Test reserve adequacy against a modelled emergency requirement, and state the answer publicly
Philippine buffer stock is 7–15 days
PUBLIC
Against a 30-day resilience benchmark, the most import-exposed large market in the region enters the El Niño peak with roughly a third of the recommended cover.Pre-position, or accept a compressed decision window in Q4
Three major exporters have restricted before
MODELLED
India, Viet Nam and Thailand have each curbed shipments during past supply shocks. A restriction is a policy event, and policy events propagate faster than harvests.Agree the trade-restraint position in advance of the shock, not during it
AFSIM produces decision intelligence. It does not make sovereign decisions, and nothing in this column is a recommendation to any government.

Regional Score Composition

Dimension profile

Indicator status across the register

A smaller number of real data points is worth more than a large number of invented ones. Data gaps are shown, not filled.

The Layering

NATIONAL REFERENCE

Singapore AIC³

One import-dependent state, 180-plus source countries, eleven food categories, a working kernel and a documented disruption history. The national implementation that proves the method.

REGIONAL OPERATIONS

ASEAN Collective C3

Interdependency, exportable surplus, correlated exposure, simultaneous substitution, reserves and coordinated response. The operating engine, scaled from one nation to a food system.

REGIONAL MEASUREMENT

AFSIM

The index, monitor, early warning, situation reporting and benchmarking layer above the engine. It is what makes the engine's verdicts comparable, traceable and trackable over time.

AFSIM COMMAND › ASEAN FOOD SECURITY INDEX

ASEAN Food Security Index

A transparent index architecture rather than a black-box score. Five dimensions, each built from named indicators with a source, a retrieval date, a normalisation band and a confidence class. Every score below is computed in the browser from the AFSIM Data Register — change the register and the index changes, because the index has no hard-coded value anywhere in it.

DIMENSION 01 — AVAILABILITY
Production, yield, trade position, exportable surplus
DIMENSION 02 — ACCESS
Prices, volatility, import cost, affordability pressure
DIMENSION 03 — STABILITY
Climate, disease, concentration, trade policy exposure
DIMENSION 04 — RESILIENCE
Reserves, diversification, substitution headroom, recovery
DIMENSION 05 — UPSTREAM
Fertiliser, feed, energy, freight — the causes behind the crisis

Every Indicator in the Index

IDIndicatorValuePeriodStatusScoreSource
VERIFIED Authoritative, or derived arithmetically from authoritative values
PUBLIC Published by a reputable public institution
MODELLED Estimated or scored under stated assumptions
PARTIAL Incomplete coverage across states or commodities
DATA GAP No reliable public value — excluded from the score
Click any indicator name to open its provenance card: source, URL, retrieval date, update frequency, normalisation band, coverage and caveats.

Why Five Dimensions, and Why Upstream Is One of Them

The first four are conventional

Availability, access, stability and resilience follow the established architecture of food security measurement, which makes AFSIM legible to institutions that already work with those categories and comparable against existing national frameworks.

The fifth is the differentiator

Fertiliser, feed, energy and freight are not food, and most food security indices exclude them. They are in AFSIM because they are where a regional food crisis actually begins. An index that reads only the food layer detects the crisis at the harvest; an index that reads the input layer detects it one to three growing seasons earlier.

Index Limitations, Stated Plainly

Normalisation bands are prototype parametersEach indicator is mapped to 0–100 using a stated band. Those bands are analytical choices made for this demonstrator, published in the Methodology view, and would require expert calibration and peer review before the index carried any weight in policy.
Equal dimension weighting is a placeholderThe five dimensions are currently weighted equally because there is no defensible basis in this prototype for weighting them otherwise. Weighting is a governance decision, not a technical one.
Country coverage is unevenCountry-level scoring draws on real production and trade-position data plus a small number of country-specific reserve and policy observations. Where no country value exists the regional dimension score is applied as an explicit proxy and labelled as such on every country card.
A snapshot is not a time seriesThis build carries one verified data snapshot. Trend arrows describe the direction of the underlying published series where that direction is stated by the source, not a computed AFSIM history. The index has no back-run yet.
The index is not a ranking instrumentA single composite number conceals the multidimensional profile that actually drives policy. AFSIM leads with the dimension profile and treats the composite as a navigation aid, not a verdict on a Member State.
MONITOR › MEMBER STATE COMPARISON

Member State Comparison

Eleven Member States against the same five dimensions. The purpose is not to rank governments but to show where the region's vulnerabilities actually sit, and to make visible how much of each country's score rests on country-specific evidence rather than a regional proxy.

The Full Comparison

Member StateAFSIMAvailAccessStabResilUpstrPaddy 2022ShareTrade positionCoverage
Paddy production is FAOSTAT 2022, retrieved 5 September 2026. Share is that state's percentage of the ASEAN paddy total, calculated from the same source. Coverage states how many of the five dimensions are backed by country-specific data rather than the regional proxy — read it before reading the score.

Country Evidence Actually Held

Member StateCountry-specific evidence in the registerConfidence
IndonesiaRice reserves at a record 5.37 Mt (May 2026); Bulog had procured 2.8 Mt by 18 May, about 70% of the 2026 target; Presidential Instruction No. 4 of 2026 mandates 4 Mt annual procurement through 2029; production rose 13.29% year on year to 34.69 Mt in 2025. Also on the ASMC drought-forecast list.PUBLIC
ThailandRoughly 60% less rainfall than normal January–April 2026; Bangkok export price US$490/t at end-June, up 27.5% year to date; DFT projects 7 Mt exports for 2026 against 2.2 Mt shipped January–April; Thai 5% broken quoted at US$420/t against Indian US$340/t. On the drought-forecast list.PUBLIC
PhilippinesNFA buffer stock sufficient for 7–15 days against a 30-day resilience benchmark; potential loss of up to 700,000 t of rice under the current event; pork import quota expanded from 54,210 t to 204,210 t in 2026 partly to offset ASF shortfalls; about 500,000 pigs vaccinated across six provinces; ASF regionalisation introduced November 2025. On the drought-forecast list.PUBLIC
Viet NamGenetic sequencing in 2026 found 83.6% of representative ASF samples were recombinant genotype I/II strains, prompting updated vaccine guidance to provincial managers in May 2026; three domestic ASF vaccines licensed since 2022. Supplies the large majority of Philippine rice imports.PUBLIC
MalaysiaWith Indonesia accounts for roughly 90% of global palm oil supply; production fell about 20% during the 2016 El Niño. On the drought-forecast list.PUBLIC
Myanmar · Timor-LesteOn the ASMC drought-forecast list. Production data held from FAOSTAT 2022; no reliable public reserve, price or policy series retrieved for this build.PARTIAL
Cambodia · Lao PDR · Singapore · BruneiProduction and trade position held from FAOSTAT 2022. No country-specific reserve, price, climate or disease series retrieved for this build — regional proxy applied on four of five dimensions.DATA GAP
MONITOR › INDEX × COMMODITY MATRIX

Member State × Commodity Matrix

The three-layer logic the platform is built on: a country's risk on a commodity is not the same thing as its dependency on another country, and neither is the same thing as that commodity's systemic importance to the region. This matrix holds the first layer. The trade matrix holds the second. The interdependency index holds the third.

SUPPLY POSITION RISK BY MEMBER STATE AND COMMODITY MODELLED
Scored by the patented risk determination kernel from the underlying supply position, import dependence, concentration and substitution availability held for each cell. Cells for which the register holds no country-commodity data are rendered as data gaps rather than as low risk — an empty cell and a safe cell are different statements, and a system that conflates them will fail quietly.

What a Cell Holds

FieldDefinitionWorked cell — Philippines · rice
Supply positionDomestic production against domestic requirementStructural deficit; largest rice importer in the region
Import dependenceShare of requirement met from tradeSubstantial and persistent PARTIAL
Import concentrationShare held by the top supplier and top threeViet Nam dominant PUBLIC
Exportable surplus available to itWhat suppliers can genuinely release, not what they produceUnder pressure — see the surplus engine
Substitution availabilityAlternative suppliers, and whether others are seeking them at the same timeThin under simultaneous demand MODELLED
Reserve coverDays of consumption held7–15 days against a 30-day benchmark PUBLIC
Kernel riskThe authoritative risk value from SG 10202109991YHIGH
Only the kernel produces the authoritative risk value in the final row. No AI model, agent or language model may originate it.
MONITOR › PRICE & MARKET MONITOR

Price & Market Monitor

Observed market conditions for the commodities that carry ASEAN food security, separated cleanly from modelled and illustrative values. Every figure on this page is a published index or quotation with a retrieval date.

FAO FOOD PRICE INDEX PUBLIC
133.3
August 2026 · up 1.9% on July, 2.5% on a year earlier · 16.8% below the March 2022 peak
FAO CEREAL PRICE INDEX PUBLIC
116.3
August 2026 · up 2.2% on July · highest since May 2024
FAO VEGETABLE OIL INDEX PUBLIC
196.9
August 2026 · third consecutive rise · highest since June 2022
FAO SUGAR PRICE INDEX PUBLIC
106.4
August 2026 · up 12% on July · highest since June 2025
FAO MEAT PRICE INDEX PUBLIC
127.9
August 2026 · up about 1% · poultry, pig and ovine meat all higher

FAO price indices, August 2026

Source: FAO Food Price Index release, August 2026 data, retrieved 5 September 2026. All five commodity groups recorded higher indices than the previous month, with markedly different magnitudes — a detail a single headline index conceals, which is why AFSIM carries the groups separately.

Rice — The Commodity That Sets the Regional Mood

THAI 5% BROKEN PUBLIC
US$420/t
Thai Rice Exporters Association quotation, May 2026
BANGKOK EXPORT PRICE PUBLIC
US$490/t
End-June 2026 · up 27.5% since the start of the year
INDIAN 5% BROKEN WHITE PUBLIC
US$340/t
May 2026 · an US$80 gap redirecting African demand
GLOBAL RICE OUTPUT 2026/27 PUBLIC
553.1Mt
FAO forecast · down 1.9% from the previous season's record
The price gap is the strategic signal, not the price level. Indian 5% broken at US$340 against Thai at US$420 has already redirected Nigerian and wider African demand toward Indian parboiled supply, compressing Thai export volumes at exactly the moment the drought reduces the volume available to compress. Thailand shipped about 2.2 Mt between January and April against a full-year projection of 7 Mt.

Upstream Markets — Fertiliser and Feed

UREA PUBLIC
US$453/t
June 2026, after peaking above US$850 in April from about US$400 pre-conflict
DAP PUBLIC
US$770/t
Up from about US$580 before the Gulf conflict
WORLD BANK FERTILISER INDEX PUBLIC
+12%
2026Q1 quarter on quarter · highest level since October 2022
FEED GRAIN DIRECTION PUBLIC
+2.5%
Maize month on month in August; sorghum +3.9%, barley +2.6%
Urea and DAP trajectory through the 2026 Gulf disruption. Sources: World Bank Commodity Markets Outlook (April 2026), WTO Secretariat data blog (10 July 2026), IFPRI/NDSU fertiliser market modelling. The June easing is real but partial: modelled trade flows return to roughly 50% of normal by August 2026 and about 75% by early 2027, so the input cost of the 2027 harvest is being set now.

Observed, Modelled, Illustrative

ClassWhat it means on this platformExample on this page
OBSERVEDA published value from a named institution, with a retrieval dateFAO Food Price Index 133.3, August 2026
MODELLEDCalculated under assumptions that are stated and can be challengedFertiliser trade recovery to ~50% of normal by August 2026
ILLUSTRATIVEExists only to demonstrate the operating model; not a statisticNo illustrative values are used on this page
MONITOR › CLIMATE MONITOR

ASEAN Climate Monitor

The climate layer is where ASEAN's food security stops being eleven national problems. A Member State can diversify its suppliers across five countries and still have all five sitting inside the same climate system — which is what is happening now.

ENSO ALERT STATUS PUBLIC
EL NIÑO
NOAA CPC advisory in force; event strengthening
VERY STRONG EVENT PROBABILITY PUBLIC
>90%
Northern hemisphere autumn and winter 2026-27, per NOAA CPC
NIÑO3.4 INDEX PUBLIC
+1.3°C
July 2026, ASMC · three-month RONI 0.9 °C for May–July
INDIAN OCEAN DIPOLE PUBLIC
POSITIVE
Neutral in July; models predict positive August–December 2026
EXPECTED EVENT TRAJECTORY PUBLIC
Niño3.4 observations and the published expectation of continued strengthening to a peak between October 2026 and January 2027. Sources: ASEAN Specialised Meteorological Centre El Niño update (August 2026), NOAA Climate Prediction Center ENSO Diagnostic Discussion (31 August 2026), IRI ENSO forecast. Points after August are the published expectation of the forecasting centres, not an AFSIM projection.

Correlated Exposure — Who Sits in the Signal

Drier and warmer than normal conditions are forecast across southern Southeast Asia. The states named in the current drought outlook are:

Indonesia Malaysia Philippines Thailand Myanmar Timor-Leste
Four of those six are among the five states that produce 90.9% of ASEAN paddy. This is the correlated exposure proposition stated as an observation rather than an argument: the region's production concentration and the region's climate signal are pointed at the same countries.

Observed Regional Effects

SignalObservationClass
Thailand rainfall deficitRoughly 60% less rainfall than normal between January and April 2026PUBLIC
Regional rice outputEstimated 2–8% decline from the compounding of El Niño and a positive Indian Ocean Dipole; the 5% midpoint is roughly equivalent to removing Thailand's entire annual export volume of 7–8 Mt from world marketsMODELLED
Philippine rice lossPotential loss of up to 700,000 tonnes under the current eventMODELLED
Global rice outputFAO forecasts 553.1 Mt for 2026/27, down 1.9% from the previous season's record, attributed in part to weather associated with El NiñoPUBLIC
Palm oil exposureIndonesia and Malaysia together account for roughly 90% of global supply; Malaysian production fell about 20% during the 2016 El NiñoPUBLIC
Haze and air qualityExposure to haze-related pollution could rise from about 25 million to about 70 million people during the peak of the eventMODELLED
Fisheries and proteinFish production, critical to Indonesian protein intake, becomes more volatile; Thailand, Viet Nam and Indonesia face multi-sector water stressPARTIAL
River basin hydrologyMekong basin conditions across five Member States — no consolidated public series retrieved for this buildDATA GAP
Gridded rainfall and drought indicesSub-national rainfall anomaly, SPI and crop-condition rasters would be ingested from meteorological and open satellite datasets in production; not retrieved for this demonstratorDATA GAP

The Decision Point Is October

The ASEAN Specialised Meteorological Centre is due to issue its October-to-December outlook through the ASEAN Climate Outlook Forum. If that bulletin confirms a strong El Niño alongside a positive Indian Ocean Dipole, the published expectation is an escalation of fire and drought emergency measures across Indonesia, Malaysia and Thailand, and a higher probability of export restrictions on rice and palm oil. A signal of rapid weakening would allow preparations to be scaled back.

This is what an early warning system is for. The decision-relevant question is not what the October bulletin will say. It is whether the substitution headroom, reserve adequacy and trade-restraint position have been quantified before it is published, because after publication the region is negotiating under a price signal rather than ahead of one.

MONITOR › DISEASE MONITOR

Animal & Plant Disease Monitor

Disease is the second correlated exposure. A single pathogen crossing borders removes protein supply from several Member States at once, and the trade measures that follow remove it from several more — which is why the propagation model reads outbreak notifications and trade restrictions as one chain rather than two.

ASF — COUNTRIES AFFECTED, ASIA-PACIFIC PUBLIC
20
Outbreaks persisting into 2026 across the region
VIET NAM RECOMBINANT STRAINS PUBLIC
83.6%
Of representative 2026 samples sequenced as genotype I/II recombinants
PHILIPPINE PIGS VACCINATED PUBLIC
~500k
Across six provinces; reported 90% efficacy in healthy ASF-negative animals
PHILIPPINE PORK IMPORT QUOTA PUBLIC
204,210t
Expanded from 54,210 t in 2026, partly to offset ASF shortfalls

African Swine Fever — Regional Position

Member StatePosition held in the registerClass
Viet NamGenetic sequencing in 2026 found 83.6% of representative samples were recombinant strains derived from both genotype I and genotype II viruses, prompting updated vaccine selection guidance to provincial animal health managers in May 2026. Three domestic ASF vaccines licensed since 2022.PUBLIC
PhilippinesNearly 500,000 pigs vaccinated in six provinces under a government-controlled programme. ASF regionalisation introduced November 2025, allowing recognition of ASF-free zones in line with WOAH standards. Animal Vaccine Development Programme launched to build domestic production capacity.PUBLIC
IndonesiaFirst batch of 120,000 doses of AVAC ASF live vaccine received in June 2025, with distribution to local authorities planned.PUBLIC
Regional frameworkASEAN ASF Prevention and Control Strategy 2023–2028, developed by a core group of WOAH, FAO, the Philippines as ASEAN lead country, Malaysia, Myanmar and the ASEAN Secretariat.PUBLIC
Live outbreak notificationsEvent-level WOAH WAHIS and FAO EMPRES-i notifications — outbreak location, affected species, status, spread and zone — would be ingested continuously in production. Not retrieved as a live feed for this demonstrator.DATA GAP

Why Regionalisation Is a Food Security Instrument

Country-level bans remove more supply than the disease does

A national-level suspension in response to an outbreak in one province withdraws that country's entire approved establishment list. Where several importing states apply the same measure to the same exporter simultaneously, the supply removed from the regional market is a multiple of the supply actually affected.

Zoning keeps unaffected supply in the market

The Philippine adoption of ASF regionalisation in November 2025, in line with WOAH standards, is the practical form of this. AFSIM tracks regionalisation adoption as a resilience indicator because it changes how much supply a given outbreak actually removes — a policy variable, not an epidemiological one.

Plant Health

IPPC phytosanitary notificationsPest outbreaks, quarantine events and plant-health restrictions across Member States. Assigned to the plant-health watch in the data spine; no live feed retrieved for this build. DATA GAP
Why it matters to the indexPlant health events act on the availability dimension through yield and on the stability dimension through the trade measures they trigger. Absent this feed, both dimensions currently under-detect a whole class of shock, and the coverage figure says so.
MONITOR › LOGISTICS MONITOR

Corridor & Logistics Monitor

Food that exists but cannot move is not available food. The logistics layer reads the corridors that carry ASEAN's food and its inputs, and states clearly where it is reading a live feed and where it is not.

Live vessel data is not connected in this demonstrator. No AIS positions, port congestion feeds or freight capacity series are running behind this page. Where a corridor status is shown, it is drawn from published reporting with a retrieval date, and it is labelled PUBLIC or MODELLED. The platform does not display a live indicator over static data.

Strait of Hormuz — The Input Corridor

STATUS PUBLIC
DISRUPTED
Closed to nearly all shipping from 28 February 2026
GLOBAL FERTILISER TRADE AT RISK PUBLIC
~⅓
Of global seaborne fertiliser trade, per IFPRI
UREA EXPORT CAPACITY BLOCKED PUBLIC
~21Mt
Annual Gulf capacity, plus about 4 Mt of DAP
MODELLED FLOW RECOVERY MODELLED
~50%
Of normal by August 2026; about 75% by early 2027
The Middle East accounts for close to a quarter of global urea exports. Iran halted ammonia production during the conflict; Qatar suspended urea, ammonia and sulfur production after damage to key facilities; India reduced urea and ammonia output on lower LNG supply. Separately, 83% of LNG transiting Hormuz in 2024 went to Asian markets. The corridor therefore reaches ASEAN food security twice — once through fertiliser cost and once through the energy cost of producing, cooling and moving food.

Regional Corridors

CorridorFood security functionLive feedClass
Malacca and Singapore StraitsThe maritime spine of ASEAN food and feed trade, and the route by which external supply reaches the maritime sub-systemNot connectedDATA GAP
Strait of HormuzFertiliser, ammonia, sulfur and LNG — the upstream inputs behind the harvestNot connected; published reporting heldPUBLIC
Johor–Singapore land crossingsPerishables, live animals, eggs and vegetables on a daily cycle with almost no bufferNot connectedDATA GAP
Mekong cross-border road corridorsRice, maize, cassava and livestock movement across the mainland sub-systemNot connectedDATA GAP
Indonesian and Philippine inter-island shippingDomestic distribution — the point at which national availability becomes local accessNot connectedDATA GAP
Five of six corridors carry no live data in this build. That is the honest state of the demonstrator, and it is a more useful thing to show a decision-maker than five green lights over static numbers.

What Production Would Connect

AIS and vessel movementCommercial or national AIS feeds for chokepoint transit counts, dwell times and route deviation
Port performanceBerth waiting times and container dwell at the region's principal food-handling ports
Freight rates and capacityDry bulk and reefer capacity as a cost and feasibility input to the substitution engine
Border and crossing statusLand crossing throughput and closure events, particularly on the perishables corridors
Cold chain integrityPrivate logistics telemetry, contributed under the ecosystem's aggregation thresholds
Until these are connected, the substitution engine can identify an alternative supplier but cannot fully price the feasibility of reaching it. That limitation is carried in the resilience dimension's coverage figure rather than hidden inside a score.
EARLY WARNING › RISK PROPAGATION

Risk Propagation

An alert on its own tells a decision-maker that something happened. A propagation chain tells them what it will do, to whom, in what order, and how long they have. This is the difference between a monitoring system and a decision intelligence system.

STEP 01 · OBSERVED PUBLIC
El Niño strengthening in the tropical Pacific
NOAA CPC advisory in force, greater than 90% chance of a very strong event through the northern autumn and winter; ASMC Niño3.4 at 1.3 °C in July 2026; positive Indian Ocean Dipole expected August–December.
STEP 02 · OBSERVED PUBLIC
Rainfall anomaly across southern Southeast Asia
Drier and warmer than normal conditions forecast for Indonesia, Malaysia, the Philippines, Thailand, Myanmar and Timor-Leste. Thailand already recorded roughly 60% less rainfall than normal between January and April.
STEP 03 · MODELLED MODELLED
Rice production risk in the concentrated producer set
Estimated 2–8% regional output decline. Four of the six states in the drought outlook are among the five that produce 90.9% of ASEAN paddy — so the climate signal and the production concentration coincide rather than offset.
STEP 04 · MODELLED MODELLED
Exportable surplus falls faster than production
Domestic requirement and reserve obligations are met before exports are released, so a 5% production fall removes a disproportionately larger share of the surplus genuinely available to importers. The 5% midpoint is roughly equivalent to removing Thailand's whole 7–8 Mt export programme.
STEP 05 · MODELLED — POLICY EVENT MODELLED
Export restriction probability rises
India, Viet Nam and Thailand have each curbed rice shipments during past supply shocks. A policy event propagates in days where a harvest propagates in months, which is why this is the step that compresses the decision window.
STEP 06 · MODELLED MODELLED
Import exposure crystallises in the deficit states
The Philippines, Malaysia, Singapore, Brunei and Timor-Leste carry structural rice deficits. The Philippines enters the window with 7–15 days of buffer stock against a 30-day benchmark and a modelled loss of up to 700,000 tonnes.
STEP 07 · THE SIGNATURE INFERENCE MODELLED
Simultaneous substitution demand exceeds regional headroom
Several deficit states seek the same alternate supply at the same moment. Each national decision is individually rational; collectively they may be impossible. Simultaneous substitute demand against regional available supply gives the headroom figure — currently thin on rice.
STEP 08 · TESTED AGAINST RESERVES VERIFIED
Reserve coverage is tested and found to be 1.45 days
APTERR earmarked stocks total 787,000 tonnes — 87,000 from ASEAN Member States and 700,000 from the Plus Three. Against combined Southeast and East Asian consumption of about 542,000 tonnes a day, the mechanism covers roughly a day and a half.
STEP 09 · KERNEL AUTHORITATIVE
Risk determination and AFSIM score movement
The patented kernel determines the authoritative risk state from the evidence chain, and the index moves. The response options that follow are generated for consideration by authorised decision-makers — the platform never issues them.

Propagation Timing

Indicative detection lead against consequence severity for each step in the chain. The steps furthest to the right are the ones AFSIM exists to surface, because they are the ones where coordination is still possible. Timings are modelled for this demonstrator.

The Same Chain, Other Triggers

UPSTREAM TRIGGER

Fertiliser dislocation

Corridor closure → input cost and availability → planting decisions → yield → exportable surplus → the same substitution and reserve steps, one to three growing seasons later. The longest lead time in the system, and the one most often ignored.

DISEASE TRIGGER

Transboundary animal disease

Outbreak notification → national import suspensions → approved establishment lists withdrawn → protein supply removed from several importers simultaneously → substitution demand concentrates on the remaining approved sources.

CORRIDOR TRIGGER

Route or crossing failure

Chokepoint or land crossing disruption → perishables first, staples later → local availability fails while national availability holds → the failure is a logistics event that presents as a supply event.

REPORTING › ASEAN FOOD SECURITY SITUATION REPORT

ASEAN Food Security Situation Report

The monthly and quarterly reporting output the concept note calls for, generated from the data register rather than written by hand. Every figure in the generated report is pulled live from the same indicators that drive the index, so the report and the dashboard can never disagree with each other.

REPORT GENERATOR

Generate the current situation report

The report is assembled at the moment you press the button, from whatever the register holds. Nothing in it is pre-written prose with the numbers dropped in.

Reporting Cadence

MONTHLY

Situation Report

Index movement, country and commodity movements, active warnings, climate, disease, trade, price, logistics and reserve position.

QUARTERLY

Regional Outlook

Seasonal outlook, scenario set for the coming quarter, structural vulnerabilities and the areas where the register's coverage most needs to improve.

ANNUAL

State of Food Security in ASEAN

Full-year index history, benchmarking across Member States, progress against ASEAN food security objectives and a published methodology revision.

EVENT-DRIVEN

Flash Update

Issued when an indicator crosses a defined threshold. The trigger set is published in the methodology so that a reader can predict when a flash update would fire.

Reporting Discipline

Generated, not composedThe narrative is assembled from the register at generation time. If an indicator is a data gap, the report says so in the position where the number would have been.
Every figure retains its classObserved, derived, modelled and data-gap classes survive into the report. A reader of the PDF can tell the same things a reader of the dashboard can.
No sovereign recommendationsThe report presents areas for attention. It does not tell a Member State what to do, and it does not describe any option as an ASEAN position.
Method changes are versionedA change to a normalisation band, a weight or an indicator definition changes the index. Those changes are published with the report, because an index that quietly re-specifies itself is not a measurement instrument.
DATA & INSTITUTIONS › AFSIM DATA REGISTER

AFSIM Data Register

The full catalogue behind every number on the platform. Indicator, dimension, geography, commodity, value, unit, period, source, URL, retrieval date, update frequency, data status, confidence, coverage and caveats. This is the credibility layer: an index that cannot show its register is an opinion with a number attached to it.

INDICATORS IN REGISTER
Across five dimensions
OBSERVED OR DERIVED
From a named public source with a retrieval date
MODELLED
Scored under stated assumptions
DATA GAPS
Declared, excluded from scoring, not estimated

The Register

IDIndicatorDimensionGeography CommodityValuePeriodSource RetrievedFrequencyStatusCoverageNotes
Click any indicator name for its provenance card, including the clickable source URL and the normalisation band used to convert the value into an index score.

Sources Actually Used in This Build

SourceWhat was taken from itRetrieved

Sources Specified but Not Yet Ingested

SourceIntended roleAccess route
ASEANstatsRegional production, trade, agricultural trade, fisheries and forestry baseline. The 2022 paddy total of 195.5 Mt and the agri, forestry and fisheries trade value of US$447 bn already carried on the platform come from this source.Portal / statistical publications
AFSISThe regional food security spine — commodity monitoring, supply and utilisation, prices, food security indicators and early warning information.Portal
FAOSTAT detailed trade matrixThe bilateral cube: exporter → importer → commodity → year → value → tonnage. The backbone of the interdependency and concentration analytics.Bulk download / API
UN ComtradeHS-level bilateral trade where more recent or more granular than FAOSTAT.API
World BankInflation, exchange rates, GDP, population and logistics indicators for the access dimension and for per-capita normalisation, which is currently a declared gap.API
WITS and BACIReconciled historical bilateral trade for backtesting and network analysis.Bulk download
WOAH WAHIS · FAO EMPRES-iEvent-level animal disease notifications with location, species, status and zone, feeding the disease propagation model.Portal / API
IPPCPhytosanitary notifications, pest outbreaks and quarantine events for the plant health watch.Portal
ASMC · national meteorological servicesGridded rainfall and temperature anomaly, drought indices, crop condition and river basin hydrology.Portal / open datasets
AIS and port performance providersVessel movement, chokepoint transits, port congestion and freight capacity for the corridor model.Commercial
National customs and statistics agenciesAuthoritative monthly national trade and production — the ground truth against which the regional cube is reconciled.National portals
Listed here rather than in the register because nothing has been taken from them in this build. The distinction between a source that has been used and a source that has been named is the distinction most data-architecture slides quietly collapse.

Data Engineering Reality

Why this is a snapshotA single self-contained HTML demonstrator cannot call most of these sources directly from the browser. Cross-origin restrictions, authentication and rate limits sit in the way, and pretending otherwise would be the exact failure the no-fabrication rule exists to prevent.
What was done insteadPublished values were retrieved, verified against the issuing institution's own reporting where possible, embedded as a dated snapshot, and tagged with source and retrieval date at the point of display.
Demonstrator data layerThe embedded register, structured exactly as a production feed would deliver it, so that replacing it changes no analytics code.
Production data architectureScheduled ingestion, schema validation, reconciliation against national ground truth, an ontology mapping layer, versioned indicator definitions and a provenance record written on every stored value.
The honest labelThis build is a VERIFIED DATA SNAPSHOT — 5 SEPTEMBER 2026. It is not a live connection, and no part of the interface claims that it is.
DATA & INSTITUTIONS › AFSIM METHODOLOGY

AFSIM Methodology

How the index and the risk measures are constructed, in enough detail that a statistician could reproduce them or argue with them. Both outcomes are useful; neither is possible with a black box.

Construction

05
Composite index
The equally weighted mean of the five dimension scores. Equal weighting is a placeholder pending a governance decision on weights, and it is stated as such wherever the composite appears.
0–100
04
Dimension score
The mean of the scored indicators within that dimension. Indicators classed as a data gap are excluded from the mean and reduce the dimension's coverage figure instead of being imputed.
0–100
03
Indicator score
Each indicator is mapped onto 0–100 by a stated linear band, oriented so that 100 is always the more food-secure end. Bands are published in the register's provenance cards and are prototype parameters requiring expert calibration.
Linear band
02
Classification
Every value is classed observed, derived, modelled, partial or data gap before it is scored. The class travels with the value into every view, chart and report.
Five classes
01
Register entry
The atomic unit: indicator, dimension, geography, commodity, value, unit, period, source, URL, retrieval date, frequency, status, confidence, coverage and notes. Nothing enters the index that does not first exist as a register entry.
Foundation

Index Bands

ScoreBandInterpretation
75 – 100STRONGAvailability, access and buffers all comfortable; no dominant systemic driver active
60 – 75STABLESound overall position with identifiable but contained weaknesses
45 – 60WATCHOne or more dimensions deteriorating; monitoring intensity should rise
30 – 45ELEVATED RISKMultiple dimensions weak simultaneously; coordination value is high and the decision window is shortening
0 – 30CRITICALSevere simultaneous weakness; the region is managing a shock rather than preparing for one

Derived Calculations

MeasureFormulaCurrent value
Production concentration, CR5Σ (top five producer output) ÷ ASEAN total output90.9% DERIVED
Production concentration, CR3Σ (top three producer output) ÷ ASEAN total output68.0% DERIVED
Herfindahl–Hirschman IndexΣ (each state's percentage share)²≈1,904 DERIVED
Reserve coverage, daysEarmarked reserve tonnage ÷ daily consumption1.45 days DERIVED
Indicative exportable surplusProduction − domestic utilisation − required reserves − normal exports − other constraintsPer commodity MODELLED
Substitution headroomRegional available supply − simultaneous substitute demandThin on rice MODELLED
CR5, CR3 and the HHI are calculated from FAOSTAT 2022 paddy production for the eleven Member States. The resulting CR5 of 90.9% independently reproduces the ASEANstats statement that five states account for about 90.9% of ASEAN paddy production — a cross-source check that is worth more than either figure alone.

Country Scoring and Its Proxy Rule

AvailabilityDerived per state from FAOSTAT 2022 production and net trade position. Real for all eleven.
StabilityRegional score, adjusted where a state appears in the current published drought outlook. Real adjustment for six states.
ResilienceCountry score where a public reserve or policy value exists — currently Indonesia and the Philippines. Regional proxy elsewhere.
Access and UpstreamRegional proxy for all eleven. No country-level price, affordability, feed or fertiliser dependence series was retrieved for this build.
The proxy is always disclosedEvery country card carries a coverage figure stating how many of the five dimensions rest on country-specific evidence. A high score on low coverage is a statement about the register, not about the country.

Division of Labour Between the Components

DATA

What actually happened

Published observations with a source and a retrieval date. Data does not reason and does not forecast.

AI · PREDICTIVE MODELS

What is likely to happen

Statistical and machine learning models producing probabilistic estimates. They may not assign a risk state.

INFERENCE ENGINE

What follows from the evidence

Reasoning over the ontology to propagate a scored event through the graph. It produces chains, not verdicts.

LARGE LANGUAGE MODEL

Reading and explaining

Reads unstructured sources, summarises and explains the evidence in language. It never originates a number.

AGENTS

Defined workflows

Prepare analyses, assemble response options and stage briefings within bounded, auditable workflows.

DECISION KERNEL

The authoritative risk state

Granted Singapore patent SG 10202109991Y. The kernel alone determines the authoritative risk value. No agent, model or language model may originate or override it.

Patent scope, stated precisely. The granted patent covers the risk determination method that the kernel implements. It does not cover the AFSIM index, the register, the monitoring layer or the platform as a whole, and nothing on this platform should be read as claiming that it does.

Known Weaknesses

No time seriesOne snapshot. Trend direction is taken from the direction stated by the issuing source, not from an AFSIM history. Back-running the index over five years of FAOSTAT, Comtrade and FAO price data is the single highest-value next step.
Normalisation bands are unvalidatedThey are analytical choices made for a demonstrator. Expert calibration would change scores materially and should be expected to.
Uneven country coverageFour Member States currently rest on the regional proxy for four of five dimensions. Their scores carry correspondingly less information.
The bilateral cube is not yet realDependency, concentration and network centrality across the trade matrix remain prototype values. Assembling the actual eleven-by-eleven-by-commodity dataset from FAOSTAT and Comtrade is what converts this demonstrator into an analytical model.
No live feedsClimate, disease and logistics all read from a dated snapshot. Each is a declared gap rather than a silent one.
DATA & INSTITUTIONS › AFSIM ECOSYSTEM

AFSIM Ecosystem & Institutional Position

What AFSIM would need around it to become a working regional capability, and — more importantly — what it must never claim to be.

No organisation named on this page has agreed to participate. Every entity below is listed as a potential data or knowledge partner identified in the concept note or implied by the data architecture. None is a partner, none has been approached in connection with this demonstrator, and no endorsement of any kind is expressed or implied.

Institutional Neutrality

What AFSIM is

An analytical and decision-intelligence layer that complements existing ASEAN institutions and national systems by integrating data into a regional early warning and decision-support capability.

MeasurementMonitoringEarly warningScenario analysisDecision support

What AFSIM is not

Not centralised ASEAN control over Member States
Not a replacement for Member State authority or national systems
Not a replacement for AFSIS, ASEANstats or APTERR
Not an officially endorsed ASEAN system, and not a government platform
Not a mechanism that makes sovereign decisions of any kind

Potential Partner Ecosystem

POTENTIAL HOST

Independent not-for-profit entity

The concept note proposes an independent not-for-profit home, preferably based in an agriculture-neutral country such as Singapore. Neutrality is the operative property: a regional index hosted by a major agricultural exporter carries an unavoidable perception problem.

POTENTIAL DATA PARTNERS

Regional and international

ASEAN / ASEANstatsAFSISIRRISEARCA / CAPRIFAOUSDA / FAS
Named in the concept note as the entities whose participation the system would require.
POTENTIAL KNOWLEDGE PARTNERS

Think tanks

ISEAS – Yusof Ishak InstitutePIDSERIA
Analytical capacity, regional policy access and independent challenge to the methodology.
POTENTIAL INDUSTRY PARTNERS

Commercial ground truth

USSECCropLife AsiaYaraFood Industry Asia
Industry closes the feedback loop between what the model recommends and what buyers can actually execute — the difference between an alternative supplier and a feasible one.
POTENTIAL TECHNICAL PARTNERS

Predictive modelling and statistics

Entities expert in predictive algorithms using AI and statistics, per the concept note. In this architecture they own the model layer; they do not own the risk determination, which stays in the kernel.

DECISION AUDIENCE

Where the output goes

AMAFSOM-AMAFAFSISNational governments
Evidence to support policy formulation, resource allocation and emergency response coordination.

Indicative Establishment Cost

LineIndicative cost (USD)Note
Advisory and consultancy leadership$60,000Year 1
System modeller, one year$110,000Year 1
AI data acquisition specialist, one year$60,000Year 1
Miscellaneous services and materials$10,000Year 1
Year 1 development total$240,000Project development period of one calendar year
Supervising experts, part time$40,000Annual from Year 2
Monitoring technical secretariat, 2 FTE$100,000Annual from Year 2
Services and overheads$10,000Annual from Year 2
Annual running total$150,000From Year 2
Figures are the indicative budget set out in the AFSIM concept note, based on Singapore rates. Final development and operational costs would depend on location and hosting organisation. CONCEPT NOTE

Long-Term Benefits Claimed for the Capability

Situational awareness and transparencyA common regional picture replaces eleven separate national pictures that cannot be compared to each other
Earlier and better-informed interventionMoving from reactive crisis response toward proactive risk management by identifying threats before they become crises
Coordinated response during emergenciesA shared evidence base for the moment when several Member States are making supply decisions simultaneously
Structural vulnerabilities made visibleIdentifying where investment is required rather than where the last shock happened to land
Progress measurementTracking movement toward ASEAN food security objectives with a consistent instrument
Preparedness for future shocksSingapore in particular benefits from greater predictability on the agrifood situation across the other Member States
Summarised from the concept note prepared by Professor Paul Teng, 1 August 2026. The benefits are the concept's claims for a fully implemented capability, not claims for this demonstrator.