The Proof

In 2025, deforestation in Brazil’s Amazon fell 11 percent — the third consecutive annual decline, the lowest rate in over a decade. In the same year, Indonesia’s deforestation surged 66 percent — an eight-year high, reversing five years of decline. Both trajectories were documented by satellite, verified on the ground, and visible to anyone who looked.

Neither was an accident.

In late 2024, Indonesian President Prabowo Subianto launched a Food and Energy Estate — a national strategic project targeting food and energy self-sufficiency within four years, built on converting forest estate land into agriculture. Forestry Minister Raja Juli Antoni proposed clearing 20 million hectares of forest estate — twice the size of South Korea. Forty-three percent of the allocation is natural forest.

The mechanism is not market-driven. The food estate carries a National Strategic Project designation, which provides fast-tracked permits, overrides normal environmental review, and subordinates conservation law to a development mandate. Concessions are allocated to politically connected operators. In South Papua, the Jhonlin Group — led by palm-oil magnate Andi Syamsuddin Arsyad — is developing 2.5 million acres of rice paddies with the assistance of the Indonesian military, having purchased two thousand Chinese excavators for $250 million. Merauke Sugar Group, owned by oligarch Martias Fangiono, holds over 1.25 million acres in sugarcane concessions. In Papua alone, deforestation rose 348 percent.

Timer Manurung, executive director of Auriga Nusantara — the environmental research organization whose satellite analysis produced these figures — characterized the surge as “planned deforestation.” Eighteen percent of all national deforestation occurred inside food estate zones. The state designated the land, allocated the concessions, facilitated the clearing, and conveyed the productive assets to private operators. The deforestation is not a failure of policy. It is the policy.

Prabowo himself, at a national planning conference in December 2024, addressed the concern directly: “Oil palms are trees, right? They have leaves, right? They produce oxygen, absorb carbon dioxide. So why are we being accused?”

Indonesia had achieved five consecutive years of declining deforestation from 2017 to 2021, under a moratorium first introduced in 2011 and made permanent in 2019. What reversed the trajectory was not a natural force but a sequence of institutional decisions: the Omnibus Job Creation Law in 2020 removed the requirement for minimum forest cover in each region. The National Strategic Project classification allowed priority programs to override moratorium protections. The moratorium map itself was revised fifteen times since its introduction, with 4.5 million hectares of forest removed from protection and subsequently granted permits for palm oil, pulpwood, logging, and mining.

The trajectory reversed because the institutional architecture that produced it was dismantled.

On April 1, 2026, a different institutional architecture took effect in Brazil.

CMN Resolution 5.268, issued by Brazil’s National Monetary Council in December 2025, requires financial institutions to verify INPE satellite deforestation data before approving subsidized rural credit. Properties flagged with native vegetation loss after July 2019, exceeding legally permitted levels, cannot access financing. The rule covers approximately $53 billion in subsidized agricultural loans — roughly one-third of all rural credit in Brazil.

The mechanism converts satellite data into financial consequence. INPE’s PRODES system monitors forest cover. The Ministry of Environment maintains a registry of flagged properties. Banks check the registry before disbursing loans. A farmer whose property shows unauthorized clearing faces not a fine or a court date but the withdrawal of the financial instrument that makes commercial agriculture viable.

Andre Lima, who leads deforestation policy at Brazil’s Environment Ministry, described the design: “We turned every bank manager who handles subsidized credit into an inspector of illegal deforestation.”

A Climate Policy Initiative analysis found that 17 percent of all rural lending disbursed between 2020 and 2024 went to farms on land deforested during that period. The resolution is designed to close that channel — to make the financial system structurally incompatible with unauthorized clearing, not through enforcement after the fact but through verification before the money moves.

The design has opponents. The Confederacao da Agricultura e Pecuaria do Brasil — the national farm lobby — is coordinating with the agricultural parliamentary front to pass legislation suspending the resolution. Six legislative decree projects are in play. CNA’s core objection: PRODES detects vegetation change, not legal status — the satellite cannot distinguish between authorized and unauthorized clearing. The mechanism is ten days old and already under political siege.

Whether the architecture holds is a political question. That the architecture was built is not.

The satellite technology is the same. Both countries generate comprehensive, verified, publicly available data on forest loss. INPE’s PRODES has operated since 1988. Auriga Nusantara’s analysis uses Sentinel-2 imagery at ten-meter resolution, ground-truthed across 49,321 hectares. The information has never been the constraint.

What differs is what each country built around the information.

Indonesia built an architecture that converts state authority into clearing permission. The president announces self-sufficiency. The minister proposes 20 million hectares. The strategic project classification overrides environmental review. Concessions flow to connected operators. The military facilitates. Satellite data documents the outcome — and the documentation changes nothing, because the clearing is the objective.

Brazil built an architecture that converts satellite data into financial consequence. INPE monitors. The ministry maintains the registry. Banks verify before lending. Credit — the instrument that makes commercial agriculture possible — becomes conditional on what the satellite sees. The documentation changes something because the architecture was designed so that it would.

Manurung noted the convergence: “Brazil’s deforestation, concentrated in the Amazon, is declining. Meanwhile, Indonesia’s is increasing. So it’s possible Indonesia could become the world’s top deforester among tropical countries in 2025.” In absolute terms, Brazil’s Amazon still lost more forest in 2025 — 579,600 hectares to Indonesia’s 433,751. But the trajectories are opposite, and the institutional architectures producing them are opposite.

A methodological caveat: INPE’s PRODES monitors primary forest in the Amazon. Auriga’s system captures loss of both primary and secondary forest across all Indonesian biomes, including young regrowth. The numbers are not perfectly comparable. The directional comparison — one declining, one surging — is robust regardless of the measurement boundary.

Both countries have the proof. Both have had it for years.

The proof was never the problem. The problem was always whether someone designed a mechanism that makes the proof matter.

Sources