# Drain the Swamp and Rice Will Grow? How Indonesia's Million-Hectare Rice Plan Ended in Smoke Source: Realpha Blog (blog.getrealpha.com) Original article and charts: https://blog.getrealpha.com/en/blog/asianometry-2026-11-12-indonesia-s-mega-rice-failure/ > Asianometry's 12 Nov 2026 episode revisits Indonesia's 1995 Mega Rice Project: 1.4 million hectares, 61% of it peatland, ending in the largest single emission event in Southeast Asian history. Listening notes on transplant conditions, irreversible steps and who gets to pull the brake. Educational, not investment advice. Published: 2026-11-13 Locale: en Tags: Indonesia, food security, agricultural policy, peatland, risk management TL;DR: Indonesia drained Borneo's peatlands to grow rice, and draining was the one step you cannot undo. The next year, over 500,000 hectares inside the project area burned. ![A straight drainage canal cuts through blackened peatland and recedes toward a vanishing point, white smoke seeping from cracks in the ground, a stilt house and charred tree trunks in the far distance, the sun pressed orange by haze](/covers/asianometry-2026-11-12-indonesia-s-mega-rice-failure-cover.png) > He who governs a state well must first remove its five calamities. > > —— Guanzi, "Measuring the Land" (Warring States period; translation mine) Asianometry's 12 November 2026 episode looks back at Indonesia's Mega Rice Project. In December 1995, President Suharto designated 1.4 million hectares of Central Kalimantan — 61% of it peatland — to be turned into a million hectares of rice paddy, even though Indonesia's own Soil Research Institute of Bogor had labelled that land "problem soils" back in 1974. Construction began in January 1996 before the cost-benefit analyses and environmental impact assessments were finished; the following year fires took over 500,000 hectares inside the project area, and roughly 10 million hectares once you count the land beyond it. The project was cancelled in 1999 after $260–500 million was spent, depending on the source, with the paddy never built. ## Java was the rice bowl — why go all the way to Kalimantan? Because Java had run out of room. Volcanic ash makes Java's soil fertile, but the island is only 9% of Indonesia's land area, and in the 1960s that 9% supported close to two-thirds of the population. Java had about 2.9 million hectares of irrigated rice land in 1905; a quarter of a century later it had reached 3.3 million, a 14% expansion. ![Two horizontal bars of very different length: the top one is filled only one-ninth of the way, showing that Java holds 9% of Indonesia's land area, while the bottom one is filled almost two-thirds of the way, showing its share of the population; at lower right, two bars of almost equal height show that irrigated rice land grew only 14% over twenty-five years.](/figures/java-land-vs-people-en.svg) Rice also mattered more than food usually does. In the inflation-wracked years after independence, the government paid its bureaucracy and army in rice, and rice was the first thing scarce foreign exchange went to. There is a saying in Indonesia: you haven't truly eaten if you haven't had rice. Between 1960 and 1964 the country imported an average of a million metric tons a year, making it the world's largest rice importer, at over twice the volume of second-place India. Suharto, who took power in 1967, had good timing. The Green Revolution's high-yielding strains IR8 and IR5 were arriving; the government subsidised fertiliser, extended cheap credit and set up BULOG to put a price floor under farmers. Then the 1973 October War and the Middle East embargo sent oil prices to the moon, lifting Indonesia's annual oil revenues past $5 billion, which paid for fertiliser subsidies, canals, roads and bridges. National output climbed from 12 million tons in 1969 to 22 million in 1981, and in 1984 the country declared rice self-sufficiency. The press called it a rice miracle, and Suharto carried the "father of development" reputation abroad on the back of it. The 1990s undid that. Output growth that decade ran at 0.5% to 1% a year while the population, over 179 million by 1990, kept growing 1% to 2%. Large capital projects in the late 1980s converted around 20% of Java's rich farmland into urban area, and then severe flood and drought damaged the harvest. In 1992 Indonesia returned to the rice market for the first time in nearly a decade; by 1995 it was importing three million tons. Java could not produce new farmland, so Suharto looked to Kalimantan — the 73% of Borneo that belongs to Indonesia. Central Kalimantan is the largest and least populated of its provinces, a frontier of forest and rivers. ## Peat looks black and rich — so why won't rice grow in it? Because that black layer is partially decomposed vegetation sitting in water, and it does not behave like soil. Plants grow on the peat, die, and add their bodies to it; over time large peat domes build up between the rivers. Fertility depends on where you stand. Coastal peat layers are shallow, roughly 25 to 100 centimetres, and enriched by the ocean, so crops are possible. Inland is another world: some peat domes hold deposits 17 to 20 metres deep, with nothing but quartz sand and granite underneath. The soils are acidic, waterlogged, short of copper, magnesium and potassium, and under some conditions they turn toxic. The climate does not help either — most rain falls in the November-to-April wet season, and the June-to-September dry season brings sustained droughts that dry the peat into prime fuel. ![On a single metre scale, the coastal peat on the left is only a thin flat sliver, while the inland peat on the right is a massive block rising twenty metres, a height difference of roughly twentyfold.](/figures/peat-depth-coast-vs-inland-en.svg) The number that stayed with me is a different one. Peatlands are 3% of the world's land surface, yet they hold about as much carbon as all the world's biomass and the atmosphere combined, locked away for thousands of years. Draining them pulls the pin. The indigenous Dayak had mapped the boundary long ago. They live in stilt houses along the waterways, dozens of families to a longhouse, eating cassava, pigs and river fish, and growing rice for themselves. Their traditional method, the ladang, starts by picking a plot according to their own seasonal forecast, clearing shrubs and small trees, running a low-grade burn strictly inside the field boundary, then sowing rice by seed. The burn keeps weeds down for about a year; after that they switch to rubber or pineapple, or leave the plot fallow for next time. What matters is the selection — only the relatively fertile shallow coastal peat, and only enough rice to eat. Even so, rainfall there is volatile enough that crops often suffer, which is why they favour fast-maturing varieties. ![On the left, only three cells out of a hundred are filled in, standing for 3% of the world's land; on the right, two bars of equal height compare the carbon held in peatlands with the carbon in all biomass plus the atmosphere combined.](/figures/peat-three-percent-carbon-en.svg) That inland peat cannot carry mass paddy cultivation was known locally and in Indonesian science: the Soil Research Institute of Bogor attached the "problem soils" label in 1974. ## Locals and scientists both said no, so why did it break ground? Because the man who decided had seen a different piece of land succeed. In June 1995, a Singapore-based Chinese company reported growing coconuts, oil palms and some rice on coastal peatland in southeast Sumatra off Jambi Province, using dykes and canals to drain the soil and keep tidal flooding out. Suharto visited, came away impressed, and his government teamed up with the company to do the same thing in Central Kalimantan. The Sumatran peat, though, was older, thinner and held less undecomposed wet material than the thick domes of Central Kalimantan; the two shared a label and little else. People said so at the time, and it went nowhere. The ageing Suharto had been in power for decades, open defiance carried political consequences, and dissent from local experts was largely suppressed. The one voice that got out was WALHI, the Indonesian Forum on the Environment, which broadcast the coming ecological damage and the threat to the area's endangered orangutans, without stopping the plan. Suharto wanted a first harvest in 1997, so construction started in January 1996 with the cost-benefit analyses, engineering studies and environmental impact assessments still unfinished. Indonesian firms and international funders were reportedly reluctant to join; Salim Group, Suharto's frequent corporate collaborator, allegedly signed on because declining would have cost it other government contracts. The build order tells you where the priorities sat. Canals went in fast, housing did not. The plan called for 20,000 homes in two years for migrants arriving from other islands; ten months in, 600 stood. It envisioned relocating 316,000 farmer families; 15,000 made the move. In April 1997 Suharto came to inspect, and for the visit the site got a helipad and freshly replanted rice seedlings so the land would look lush. He stayed two hours, met a few people, and declared the project a success. ![Two pairs of horizontal bars: the promised bar runs almost the full width of the figure while the delivered bar is cut down to a short stub, and the same lopsided ratio repeats for houses and for farm families.](/figures/promised-vs-delivered-en.svg) ## How does clearing farmland become the region's largest emission event? Through the canals. The thousands of kilometres of gravity-fed channels meant to carry river water to the paddies were poorly planned throughout, and worst of all they cut straight through the peat domes between the rivers, systematically draining the water out of the peat. Forest went at the same time: over 50,000 hectares cleared inside the project area, not counting illegal logging beyond it. ![Two panels side by side: on the left a peat dome between two rivers sits soaked under a high water table, and on the right the same dome is sliced through by a vertical canal, the water line has dropped to the base, and the crown of the dome has cracked open.](/figures/canal-cuts-the-peat-dome-en.svg) Lost canopy, dried peat and illegal burning did the rest. Satellites picked up fires as early as January 1997. That year, more than 500,000 hectares of the designated 1.4 million burned, some 80% of it peatland, with an average of 51 centimetres of peat releasing carbon banked over millennia. Fires kept breaking out until the rains finally stopped them in June 1998. Counting beyond the project boundary, an estimated 10 million hectares burned, blanketing the region in a haze that forced daily life to a halt — and once the smoke reached Singapore and Malaysia, outsiders started asking what had been done in Central Kalimantan. The 1997 fires rank as one of the single largest peak emission events in Southeast Asian history, and drove one of the largest annual increases in atmospheric carbon since records began in 1957. The end came quickly. The 1997 Asian Financial Crisis crushed the rupiah by some 80%, inflation climbed, students filled the streets again, and Suharto resigned in May 1998 in favour of Vice President BJ Habibie. The World Bank sent an engineer to investigate, and his 1998 report set out the lack of planning, the disastrous results and the economic damage. The new president cancelled the project in 1999. Against a 1996 projection of $2–3 billion, final spending is unclear, somewhere between $260 and $500 million depending on the source, of which $200 million came out of Indonesia's Reforestation Fund — reforestation money spent on a generation-defining round of deforestation. A sizeable share went to corruption and fraudulent subcontracts. ![Three squares share one corner: the smallest is the 500,000 hectares that burned, the middle one is the 1.4 million hectares of the project area, and the much larger outer square is the roughly 10 million hectares burned once land beyond the project is counted.](/figures/burned-area-nested-squares-en.svg) The human ledger deserves its own line. The Dayak were promised investment and prosperity; they got burned croplands, fish ponds disrupted by canals, and traditional land taken over by new arrivals. The 15,000 families who uprooted themselves for the frontier found a denuded wasteland with few basic services and poor infrastructure. After the project ended, most of the land was abandoned or turned to oil palm, and it still burns readily. A 2011 study led by Agata Hoscilo examined a land block inside the project area and found that over half of it had burned at least once in the six years after the project ended, with half of that burning a second time. Stopping the fires means stopping the drainage, bringing the canopy back and restoring the peat, so the work now is blocking canals and revegetating with local species while checking that they survive. President Jokowi started the BRG peatland restoration programme in 2016, and in 2020 tentatively began a food estate programme to plant rice on 165,000 hectares of partially rehabilitated former project land, continuing under President Prabowo. Rice has been grown on a thousand or so hectares, alongside a few setbacks. ## So what can I do with this? The usable part of this episode is how to read the sentence "our pilot worked, we're rolling it out to ten more sites next year." You hear it on earnings calls, in trade press, in a friend's business plan — and Sumatra versus Central Kalimantan spells out the condition it rests on. I run three questions against it. First, transplant conditions: what differs between the demonstration site and the place being copied? Sumatran peat was old, thin and drier; the Central Kalimantan domes run past 17 metres. When a company talks about replication, I go looking for that matching difference — a different customer mix, a different power price, a different regulator, a different yield baseline. The closer the difference sits to physics, the less effort can fix it. Second, reversibility: if this is wrong, can it be undone? A wrong machine can be sold at a discount; peat water, once drained and burned off into the sky, is gone, and the canals take twenty years to plug. So I lay out the steps of a plan and ask which one is irreversible, then look at where that step sits in the sequence. Mega Rice put the most irreversible step, drainage, first, and the environmental assessment after the groundbreaking. ![Two timelines are stacked for comparison: the upper one puts the solid-marked, irreversible drainage step at the far left and the assessments at the far right, while the lower one moves that irreversible marker to the far right.](/figures/irreversible-step-order-en.svg) Third, the brake: who in this project is allowed to say stop? In Central Kalimantan the technical objections were suppressed and WALHI had to broadcast from outside the system, which did not hold. In a company, I want to know whether anyone's job is to say no and keep their job afterwards. Financial statements won't tell you; who answers the bad-news question on the call, and how, tells you something. I run these on my own holdings too. Strip the nouns away and none of the three is novel; the hard part is that a successful demonstration makes you not want to ask — and that April helipad with its replanted seedlings was built to produce exactly that feeling. ## Further reading - Asianometry, "Indonesia's Mega Rice Failure," 12 November 2026 - The Soil Research Institute of Bogor's 1974 assessment labelling Central Kalimantan peatlands "problem soils" - The World Bank's 1998 report following its engineer's investigation of the Mega Rice Project - Agata Hoscilo and colleagues' 2011 study of fire frequency inside the project area - Indonesia's Peatland Restoration Agency (BRG, founded 2016) and the food estate programme on former project land from 2020 ## The One Thing to Take Away Put the irreversible step last. Every plan has moves you can walk back and one or two you cannot: draining a peatland, felling thirty-year forest, signing a ten-year exclusive, putting your savings in at once. Mega Rice ordered the least reversible step first and left the reversible assessment for later. Here's something I've tried, offered rather than assigned: take one thing you're about to copy from someone else — their parenting, their cutting diet, their career switch, their spreadsheet — and write down three conditions they have that you don't. Three, and you know what to adjust; fewer than three, and you haven't understood the case yet, so hold off on the step you can't walk back.