BGN Asks the Question It Should Have Asked First: Is 63 Million Beneficiaries Real?

MBG Watch · 2026-07-23

The premise

On July 21, 2026, the Deputy Head of the National Nutrition Agency (BGN), Agustina Arumsari, stood at the Presidential Palace complex and said the sentence that should have been asked two years ago: "Jadi yang pertama kami perbaiki, 63 juta itu beneran nggak sih? Itu pertama." (So the first thing we fix: is 63 million even real? That's first.)

She said it plainly, in the language people actually use. Not "we are conducting a data integrity verification exercise." The question itself is the news. The program's implementing agency just admitted it does not know whether its own headline number — the denominator for every budget line, every SPPG contract, every per-meal allocation — corresponds to actual children.

What the evidence supports

The 63 million figure (sometimes cited as 62.7 million) has been BGN's working number since the national rollout. It is not the program's target. The target, set in the 2025 budget documents and repeated in presidential directives, is 82.9 million beneficiaries. The gap — nearly 20 million children — was already documented in MBG Watch's "Equity of Access" analysis: rural, remote, and administratively invisible populations systematically excluded by a delivery model that requires a registered school or SPPG within reach.

What Arumsari confirmed is that the 63 million itself is a ceiling, not a verified count. "Kalau itu berapa yang bener, lalu berapa yang seharusnya kami berikan? Daerah-daerah mana, anak-anak yang mana, yang memang harus kami dahulukan" (If that's how many are real, then how many should we actually serve? Which regions, which children, should we prioritize?). The validation exercise she announced is the first time BGN will attempt to answer that.

The 30-day evaluation she described — ordered by the new BGN leadership under the President's July directive — has two stated priorities: beneficiary data validation and restructuring SPPG incentive schemes. The incentive reform is the tell. Arumsari noted that SPPG incentives have been "disamaratakan" (uniformly applied) despite wildly different service capacities and coverage areas. A kitchen serving 500 meals 300 meters from a school gets the same per-meal incentive as one serving 400 meals 500 meters away. The perverse incentive is built in: inflate the count, claim the flat rate.

What the evidence does not support

The official line is that this is a routine "pembenahan" (improvement) exercise. That framing does not survive contact with the arithmetic.

First, the corruption machinery documented across MBG Watch's "One Failure, Not Four," "The Ompreng Toll," and "Why MBG Has No Measurable Nutrition Outcomes" runs on inflated beneficiary counts. The three extraction clusters — SPPG approval inflation, ompreng kickbacks, and the electric motorcycle procurement scheme — all scale with the number of claimed beneficiaries. More beneficiaries means more SPPGs approved, more meals "served," more budget executed, more nodes to extract from. If the validated count drops from 63 million to, say, 50–55 million (a plausible range given the equity gap and duplicate-counting risks Arumsari herself flagged: "tak ada penerima ganda karena satu sekolah dilayani lebih dari satu SPPG"), every budget line derived from that denominator shrinks.

Second, the KPK's 10 governance findings from March 2026 did not specifically flag beneficiary data integrity. Either the KPK's audit scope missed the denominator problem, or BGN's new leadership is discovering a failure the anti-corruption body did not. Both are damning in different ways.

Third, the Kejagung's halted evidence collection at SPPG sites ("The Data Door Closes") means the forensic trail on how inflated counts translated into inflated contracts is now harder to reconstruct. BGN's validation may produce a cleaner number going forward, but it cannot retroactively audit the SPPGs already paid on the old number.

The least-harm path

The arithmetic BGN is not doing in public is the arithmetic the public and the DPR should demand.

At Rp10,000 per meal, a 10 million beneficiary overcount is Rp100 billion per feeding day. Over a 200-day school year, that is Rp20 trillion — a sum that exceeds the entire 2025 MBG allocation for several provinces combined. The "Fiscal Displacement" analysis projected a Rp174 trillion allocation for 2027 assuming the 82.9M target. If the real baseline is 50–55M, the 2027 projection is a fiction built on a fiction.

The SPPG incentive restructuring Arumsari promised is necessary but not sufficient. A fairer per-meal incentive tied to verified meals served and distance-adjusted costs would reduce the inflation incentive. But it does not answer what happens to SPPGs that were approved, funded, and paid on beneficiary lists that BGN now admits may be substantially fictitious. Restitution? Contract termination? Blacklisting? The 30-day evaluation timeline does not include a forensic look-back.

The prioritization of high-stunting districts — per Prabowo's direction — is the one constructive signal. It means the validated data, if produced honestly, could finally steer resources to the children the program was named for. But prioritization requires a denominator. If the denominator is still speculative, the prioritization is performative.

What I'm uncertain about

• Whether the 30-day validation will use independent verification (village-level spot checks, cross-referencing with Dukcapil/Kemendikbud data) or remain an internal BGN reconciliation exercise. The latter reproduces the conflict of interest.
• Whether the Kejagung's evidence freeze will be lifted in time for the validation to cross-reference corruption findings with beneficiary lists. Without that, the validation cleans the data forward but leaves the past opaque.
• Whether DPR Commission IX and the Budget Committee will treat the validated number as a binding baseline for the 2027 budget discussion, or whether the 82.9M target will persist as a political anchor regardless of what the count reveals.
• The exact mechanism for SPPG incentive restructuring. "Tidak disamaratakan" (not uniform) is a principle, not a formula. The formula determines whether the corruption incentive is actually broken or just reshaped.

The question BGN's admission forces is not "how do we fix the data?" It is: "What did we spend, who did we pay, and what did they deliver — based on a number we now admit we never verified?"

The question that follows from the arithmetic

If the validated count comes in at 52 million, the Rp10,000 per-meal allocation for 63 million beneficiaries was always a transfer to the delivery chain, not to children. The corruption defense — "the program is working as designed" — collapses when the design's central variable is acknowledged as unverified. The DPR's next hearing should not ask for a reform plan. It should ask for the reconciliation: every SPPG contract, every disbursement, every incentive payment, mapped to the beneficiary list BGN now says it cannot trust. The arithmetic is simple. The politics of answering it are not.