ASEAN's Rice Yield Gap Is Real. The Spending Isn't Aimed at It.
HONG KONG, Nov 20, 2026 - (ACN Newswire) - Rice yields across Southeast Asia sit near half of tested potential, while 2026 budgets keep paying for area, stocks and cheap rice.Indonesia's rice output rose 13.29% in 2025, to 34.69 million tonnes. Very little of that came from yield.The country ha

HONG KONG, Nov 20, 2026 - (ACN Newswire) - Rice yields across Southeast Asia sit near half of tested potential, while 2026 budgets keep paying for area, stocks and cheap rice.

Indonesia's rice output rose 13.29% in 2025, to 34.69 million tonnes. Very little of that came from yield.
The country harvested 11.32 million hectares of paddy, against 10.05 million a year earlier. That is 12.6% more land. Set the two changes side by side and output per hectare moved about 0.6%. Take the land gain out of the total and roughly 5% of the increase survives.
So the biggest rice producer in the region did what the region has done for a decade: it found the extra tonnes in more fields, not in fuller ones. That choice has a price. Land that can be flooded and drained is finite, and Indonesia is already pulling back from the cheapest way to get it, which was clearing forest.
The rest of this piece is about the other variable, the one that could double the harvest without moving a fence: yield per hectare.
Where the extra rice actually came from
The official numbers make the split easy to see. Indonesia's statistics agency put 2025 paddy output at 34.69 million tonnes, up 4.07 million from 30.62 million. Harvested area went from 10.05 million hectares to 11.32 million.
Multiply the old yield by the new area and the land accounts for 3.87 million of those 4.07 million tonnes. That is 95% of the gain. Yield supplied the remaining 0.2 million tonnes, or about 5%.
Indonesia | 2024 | 2025 | Change |
Harvested paddy area | 10.05 million ha | 11.32 million ha | +12.6% |
Paddy output | 30.62 million t | 34.69 million t | +13.29% |
Output per hectare, same basis | index 100 | index 100.6 | +0.6% |
Tonnes from area | — | 3.87 million t | 95% of the gain |
Tonnes from yield | — | 0.2 million t | 5% of the gain |
Source: ANTARA, citing Statistics Indonesia and the agriculture ministry, January 2026.
The stock numbers moved in step with the area numbers. State logistics agency Bulog bought 4.11 million tonnes of paddy from farmers by October 2026, past its full-year target of 4 million. The national reserve peaked near 5.4 million tonnes, the highest since independence, and stood at 5.25 million on 30 July.
A reserve of that size is a real achievement. It is also a purchase order. At the floor price of Rp6,500 a kilogram, buying 4.11 million tonnes put public money into farmers' hands without asking any of them to grow more per hectare.
Half a harvest, on ISEAS's own estimate
The number that frames the whole argument comes from a Singapore think tank, not from a statistical office. A FULCRUM commentary published by ISEAS in May 2026 states that rice yields in the region are only about half of what they can be.
If that holds, the region could double its rice on the same land. Nothing in that claim requires a new variety from a laboratory. Vietnam already farms at 6 tonnes per hectare, roughly double what the regional average implies, on the same tropical soils and under the same monsoon.
ISEAS lists three routes to close the gap: shared agricultural technology platforms, supportive policy, and digital precision agriculture. All three are delivered farm by farm, field by field. None of them arrives as a shipment a state logistics agency can book.
That is why the gap survives. A reserve can be ordered in a single procurement. A yield gain has to be earned on several million smallholdings, each deciding separately whether the extra fertiliser is worth the money this season.
The estimate is an estimate. It comes from a policy commentary, not from a measured census of every plot, and it carries the credibility of the institution rather than the precision of a survey. What makes it usable is the direction and the size: every serious appraisal of the region's rice sector puts the gap in the tens of percent, not in the low single digits.
One average hides two farms. An irrigated field with a reliable pump and a rainfed field that waits for the monsoon do not share a ceiling, and the same seed, fertiliser and labour produce different harvests on them. Any single regional figure blends the two, which is why a district that looks close to its potential and a district that is far from it can report the same number of tonnes.
That matters for policy because the two farms need different things. The irrigated field needs better management of water it already has. The rainfed field needs water it does not have, or a shorter-season variety that can finish before the rain stops. The first is a service business. The second is capital investment, and capital does not arrive farm by farm.
What a full hectare requires
Yield is not one decision. It is four, and each one is local.
Water has to arrive on time and leave when it should, which on a delta field means levelling and drainage rather than rainfall. Plants have to be spaced so that each one has room to tiller, which is why seeding rates matter more than seed quality. Nutrients have to arrive when the plant can use them, not when the subsidy is distributed. Pests have to be caught before they spread, which needs scouting rather than calendar spraying.
None of these can be shipped as a product. A bag of urea can be sold to any farmer in any province. Levelling, drainage, seed rate and scouting depend on the slope, soil and water source of the plot in front of you, so they arrive as a service, delivered by someone who has to be there.
The Vietnamese programmes show what the service does when it reaches the field. In Vinh Long, drone spraying cut pesticide and fertiliser use by 20% to 30%, and some models raised yields by 15% to 20%. Ca Mau has registered nearly 55,000 hectares across 20 communes under the one-million-hectare scheme, with pilots of 100 to 500 hectares and at least 20 trained local pilots per site.
Read the last clause again. The binding constraint in that pilot is not the drone. It is the twenty people who can fly it, keep the records and read the map. Machinery moves faster than trained operators do, and that is why service programmes spread at the speed of training.
The instruments are built for shortage
Look at what the region actually operates, and the shape of the problem becomes clear. Almost every regional instrument is a stock, a price, or an emergency line. Almost none of them is a yield programme.
The ASEAN Plus Three Emergency Rice Reserve holds 787,000 tonnes of earmarked rice. Member states contribute 87,000 tonnes, or 11%. The three northeast Asian partners contribute 700,000 tonnes, or 89%.
In 2024 the reserve's Tier 3 mechanism released 5,622 tonnes. That is 0.71% of the stock.
APTERR, the regional rice reserve | Value |
Earmarked reserve | 787,000 t |
Contributed by ASEAN members | 87,000 t (11%) |
Contributed by Plus Three partners | 700,000 t (89%) |
Released under Tier 3 in 2024 | 5,622 t (0.71%) |
As a share of Indonesia's national reserve (5.25 million t) | 15% |
Source: ASEAN Secretariat, AMAF statement, 29 April 2026; release data compiled from APTERR reporting.
Read the table in one direction and the reserve looks large. Read it in the other and a regional pool of 787,000 tonnes is worth 15% of what one member state holds on its own. Nine tenths of the pool is pledged by countries that are not in the region, so every activation is also a diplomatic request.
The ministers' statement from April 2026 shows where the attention goes. Facing higher energy and freight costs and tighter input supply, the ASEAN Ministers on Agriculture and Forestry agreed to five measures: activate the full reserve machinery, widen the reserve from rice to edible oils, sharpen production and price monitoring through the regional food security information system, share national fertiliser and feed stocks and demand forecasts, and promote precision farming and biofertiliser with development bank guarantees behind procurement.
Four of the five are about visibility and emergency stock. The fifth, precision farming, is the only one that touches yield, and it arrived in the statement without a tonnage, a budget, or a target area.
Three budgets, one variable
Follow the money in the three largest markets and the pattern repeats. The instruments are large, they are well funded, and they act on price, stock and area.
In Indonesia the food security budget went from Rp139.4 trillion in 2025 to Rp210.4 trillion in 2026, a rise of 45.5%. The state sets a paddy floor of Rp6,500 a kilogram and buys at it regardless of quality. The self-sufficiency ratio went from 87.14% in 2024 to 110.25% in 2025, on the area expansion described above.
In the Philippines the instrument is a consumer price. The agriculture department's rice programme sells at PHP20 a kilogram to 15 million families, about 60 million people. At 6 million tonnes a year, that is roughly PHP120 billion of subsidy, and the tariff schedule on imports runs between 15% and 35%, adjusted in steps of five percentage points for every five percent move in the reference price.
In Vietnam the instrument is a field programme. The one-million-hectare low-emission rice scheme has 1,000,000 hectares as its target, and a Canada-funded project attached to it expects 48,200 direct beneficiaries and 748,000 tonnes of avoided carbon dioxide by 2030.
Instrument | Who runs it | Size | What it moves |
Paddy floor price and procurement | Indonesia, Bulog | 4.11 million t bought at Rp6,500/kg | Farm income, stock |
Food security budget | Indonesia | Rp210.4 trillion, +45.5% | Inputs, irrigation, machinery |
Cheap rice for poor households | Philippines | 6 million t a year, PHP120 billion | Consumer price |
Rice tariff band | Philippines | 15% to 35% | Import price |
Low-emission rice scheme | Vietnam | 1,000,000 ha target | Yield, cost, emissions |
Emergency reserve | APTERR | 787,000 t | Shortage |
Sources: ANTARA, August 2026; PIA, January 2026; Bureau of Customs, 2025; Vietnam.vn, 2026.
Only one row in that table moves yield. It covers one million hectares in one country, and even there the reported gains come from demonstration models rather than from the whole enrolled area.
The asymmetry is not a failure of will. Cheap rice for 60 million people is a political promise that has to be kept every week. A yield programme pays out over three or four seasons, through cooperatives, extension workers and machinery rings, and its results are visible only after the harvest. Budgets follow the weekly promise.
A floor price that pays the same for every tonne
There is a second reason the yield line stays flat, and it sits inside the price.
Bulog buys paddy at Rp6,500 a kilogram without grading for quality, according to the official account of the policy. A flat price at the farmgate pays the same for a well-managed crop and a poorly managed one. When management cannot change the price, it can only change volume, and the cheapest way to add volume is usually more land or more nitrogen.
The Philippine retail programme pulls from the other end. A PHP20 kilogram for 60 million people sets an expectation that rice will be cheap, which is a promise about the shelf, not about the field.
Both instruments settle a market at a chosen number. Neither pays for the tonne that is not there. A farmer who could add half a tonne per hectare has no buyer for that half tonne before it exists, so the investment has to be financed on a hope about next season's price.
The counter-example is already running. Vietnam's low-emission scheme pays for a labelled product, with traceability and measured emissions attached, and buyers in Japan, Australia, Singapore and Malaysia have taken cargoes under that label. The premium is small and the volume is small, but the direction matters: the payment is attached to a measured difference, not to a flat floor.
What the published numbers disagree on
Indonesia's 2026 rice crop has three official-looking numbers in circulation: 31.3 million tonnes, 34.77 million tonnes, and 38.6 million tonnes. The spread between the highest and the lowest is 23.3%.
Part of that is calendar. One figure is a harvest estimate for the calendar year, another is a forecast for the 2026/2027 marketing year that starts mid-year. Part of it is the consumption base: one set of figures implies national use of 15.4 million tonnes, another implies 31.1 million tonnes, a difference of a factor of two in the denominator.
When the numerator and the denominator can both move that much, a self-sufficiency ratio is a statement about a chosen pair of numbers. It is not a statement about how much rice is in the field.
That matters for anyone who wants to sell a yield gain. A service that promises more tonnes per hectare needs a baseline that both sides accept, and the region does not yet have one at national level. Plot-level measurement is easier than national reconciliation, which is one more reason the commercial layer is being built at the cooperative rather than at the ministry.
The weather is the honest counter-argument to everything above. If a dry year cuts the crop, output falls whether or not anybody invested in yield, and the yield story gets buried under a climate story. The April 2026 ministerial statement itself points at energy costs, freight and input supply rather than at agronomy.
Both things can be true. A cost shock and a technology gap can hit the same harvest. What separates them is the response: a cost shock is answered with stocks, subsidies and price floors, and those are exactly the instruments the region has built. A technology gap is answered with a service delivered to a farm, and that is the instrument the region has not built at scale.
The unit that can be sold
Here is where the arithmetic becomes commercial. Indonesia fixes a paddy floor of Rp6,500 a kilogram, which the agriculture ministry describes as about US$0.36. One tonne is therefore worth about $360, and the buyer is the state, at a published price, before the season starts.
A yield gain can be priced against that. A 20% lift on Vietnam's 6 tonnes per hectare is 1.2 tonnes, worth about $430 a hectare at the Indonesian floor. Whether a service can be delivered for less than that is the only question a supplier has to answer. It is a question with a number attached, which is more than can be said for a bag of urea.
This is why the interesting work in the region is not in seed sales. It is in metering. Vinh Long's rice programme runs an AI-based system for measuring, reporting and verification, built to record fertiliser use, water use and emissions plot by plot. Drone operators log flight paths, application rates and dates. Cooperatives hold the records. Once a hectare has a verifiable baseline and a verifiable harvest, a payment can be attached to the difference.
Three payers are already visible. The state pays through procurement at a fixed floor. Carbon finance pays for measured reductions, which is what the Canada-funded project's 748,000 tonnes are for. Buyers with a brand claim pay for traceability, which is what the low-emission label is for.
Selling a bag of fertiliser needs no measurement. Selling a tonne needs one, and the measurement is the part that has been missing.
What would change the number by 2030
Three markers will show whether the region moved the variable or only talked about it.
The first is Indonesian area. If harvested area expands again in 2027 while output per hectare stays flat, the model has not changed, and the forest-clearing pullback will be the binding constraint rather than a choice.
The second is the share of enrolled area that meets the full standard in Vietnam's scheme. Registration is cheap; compliance is not. A programme that reports enrolled hectares and compliant hectares separately is telling the truth, and the gap between the two is the real adoption rate.
The third is the reserve. If APTERR's stock keeps growing while its drawdown stays near 0.71%, the region is buying insurance against a shortage it is not preventing.
One test arrives with the next harvest. The 2026/2027 season runs into higher input costs and a drier pattern in parts of the region. If output falls while area holds, yield did the damage. If area falls and output holds, yield carried the load. Either result answers the question the budgets have been avoiding.
Limitations
- The "half of potential" figure is an ISEAS estimate from a policy commentary, not a measured statistic; other appraisals of the regional gap differ in method and level.
- Indonesia's 2026 production is published on at least three perimeters that do not reconcile, so year-on-year comparisons in this piece use the 2024 and 2025 series only.
Sources
- ISEAS FULCRUM: Heeding the call for collective resilience in regional food security — commentary, 1 May 2026.
- ASEAN Secretariat: AMAF Statement on Safeguarding Food Security and Agricultural Supply Chains Amid Global Uncertainties — adopted 29 April 2026.
- ANTARA: Indonesia confident on food security outlook as farm output rises — 10 August 2026.
- ANTARA: Indonesia food self-sufficiency ratio and 2025 paddy output — January 2026, citing Statistics Indonesia and the Ministry of Agriculture.
- Shangbao Indonesia: Bulog procures 4.11 million tonnes, reserve peaks near 5.4 million — 9 October 2026.
- Philippine Information Agency: Higher budget for agriculture means long-term food security — 8 January 2026.
- Bureau of Customs, Philippines: Customs Memorandum Circular 260-2025, rice tariff schedule — 2025.
- Vietnam.vn: Ca Mau registers nearly 55,000 hectares for the one-million-hectare rice scheme — 2026.
- Vietnam.vn: Vinh Long puts drones and an AI verification platform into rice production — 2026.
- Tuoi Tre News: Canada-funded project targets 48,200 beneficiaries in Vietnam's low-emission rice sector — 28 October 2026.
Copyright 2026 ACN Newswire. All rights reserved. www.acnnewswire.com
Tokyo Overnight
Japan markets, decoded. Get the brief in your inbox.
No spam. Unsubscribe anytime.