Final Assessment: What BGN’s 30-Day Kitchen Evaluation Actually Fixed — And What It Didn’t

MBG Watch · 2026-10-02

The answer in one line

BGN’s 30-day kitchen evaluation appears to have fixed part of the sanitation screen. It did not, on the public record available as of 2 October 2026, fix the operating system underneath it.

That distinction matters. A kitchen without a hygiene certificate can be closed. A kepala SPPG who violates discipline can be removed. Those are real acts, and they reduce some immediate risk. But MBG’s larger failure has never been only one dirty kitchen. It is the combination of unsafe food handling, weak verification, opaque partner selection, inflated procurement channels, disputed beneficiary numbers, and a program architecture large enough to hide mistakes until children are already sick.

The 30-day evaluation was announced in mid-July, just before the 17 August deadline set by angry SPPG partners threatening a national kitchen lockout. BGN needed to do two things at once: prevent a partner rebellion and show that the new leadership could impose order on a program already carrying food-safety and corruption alarms.

Five weeks after the window closed, the clearest judgment is this: BGN bought time, found some failures, and acted on a visible subset. It has not yet shown that it repaired the machine.

What BGN actually disclosed

The closest thing to a post-window result came not as a complete public report, but as a sequence of operational announcements.

On 27 August, RRI reported that the government would permanently close 455 SPPG that failed to meet the requirements for the Sertifikat Laik Higiene Sanitasi, or SLHS. The same disclosure said 27,485 SPPG had operated; 3,515 SPPG in 324 districts/cities still lacked SLHS; 3,060 had not yet applied; 455 failed the requirement; 27,022 had submitted BGN profiling; 463 had not submitted profiling and could face suspension; and 249 kepala SPPG had been dismissed for violations including food poisoning, phishing, absence of more than 10 working days, and other discipline breaches (RRI, 27 August 2026).

That is the strongest evidence of concrete output from the evaluation period. It shows that BGN did more than talk. It identified non-compliant kitchens, tied at least some enforcement to sanitation certification, removed some unit heads, and shifted beneficiary determination away from individual SPPG toward BGN.

But it is not a final evaluation in the accountability sense. It does not disclose:

• a full grading distribution of the roughly 27,000 kitchens;
• which SPPG were cleared, conditionally cleared, suspended, permanently closed, or awaiting correction;
• which failures were sanitation failures, document failures, infrastructure failures, cold-chain failures, payment failures, or partner-integrity failures;
• how many beneficiary routes were interrupted and how many were safely reassigned;
• whether the same kitchens later reopened, stayed closed, or shifted under new operators;
• whether the evaluation changed procurement rules, partner approvals, bank-account controls, or evidence-sharing with law enforcement.

That absence is the central finding. BGN has disclosed enforcement fragments. It has not disclosed a final public evaluation ledger.

The sanitation screen improved

The clearest area of movement is food-safety administration, especially SLHS.

This began before the window closed. On 28 July, Detik reported BGN saying 833 SPPG had been permanently closed for hygiene and sanitation problems. Sudaryono, then head of BGN, said suspended kitchens would be covered through redistribution to other kitchens, and that closures were tied to cleanliness, sanitation, wastewater treatment, and technical requirements for MBG service (Detik, 28 July 2026).

By late August, the more specific public figure was 455 SPPG failing SLHS requirements. By early September, the system was still producing additional closures. CNA Indonesia reported that BGN temporarily closed 1,999 SPPG because they had not registered for hygiene inspections required to obtain SLHS. Those were distributed across 351 in Sumatra, 1,417 in Java and Madura, and 231 in Kalimantan, Sulawesi, Bali, Nusa Tenggara, Maluku, and Papua. CNA also reported that 27,022 kitchens were distributing free meals as of 7 September (CNA Indonesia, 8 September 2026).

The September action cuts both ways. It shows that BGN kept enforcing after the initial 30-day window. It also shows that the 30-day window had not completed the sanitation baseline. If nearly 2,000 kitchens still had to be temporarily closed on 7 September for not registering for hygiene inspection, then the evaluation had not converted the network into a verified-safe kitchen system by mid-August.

The real fix here is partial:

• BGN made SLHS non-negotiable in public language.
• It attached real consequences to missing or failed certification.
• It removed some SPPG heads.
• It began to connect kitchen status to beneficiary assignment and profiling.

What it did not yet show is a live public register: every SPPG, its certification status, last inspection date, operator, beneficiary count, incident history, correction deadline, and reopening condition.

Without that register, enforcement remains episodic. The public sees numbers. It cannot see the operating map.

Food-safety outcomes did not show recovery

If the 30-day evaluation had structurally reduced risk, the first expected signal would be a visible fall in poisoning incidents. The record after the window does not show that.

Katadata reported that by 4 September, BGN had recorded 5,482 poisoning victims since Sudaryono took office on 22 July. Those victims came from 71 poisoning incidents tied to 62 SPPG. BGN told the DPR that more than 80 percent of the causes were SOP violations, and that Java and Madura accounted for 45 incidents affecting 3,799 beneficiaries, leading to 38 temporary SPPG closures (Katadata, 4 September 2026).

Kompas, citing Ministry of Health surveillance data as of 2 September, reported a much wider cumulative figure: 50,059 people affected, across 560 incidents in 245 districts/cities and 36 provinces. The Kompas report described the recurrence as systemic and quoted calls for more radical change than ordinary SOP repair (Kompas, 4 September 2026).

Those two records are not identical because they refer to different counting frames: BGN’s count since the new leadership period, and broader surveillance or cumulative counts. But they point in the same direction. The poisoning pattern continued after the evaluation period.

This is the hard test for BGN’s claim of repair. A hygiene certificate is a precondition. It is not food safety by itself. Many MBG poisoning risks sit in operating time and temperature:

• food cooked too early;
• food held too long before consumption;
• delayed distribution;
• inadequate cold storage;
• poor separation of raw and cooked ingredients;
• weak recall and incident reporting;
• head-of-kitchen absence during the critical pre-dawn production window.

BGN’s post-evaluation disclosures acknowledge this. Sudaryono identified SOP violations as the majority cause. CNA reported BGN saying poisoning was mostly caused by improper food storage and delayed distribution. That means the problem is not merely whether a kitchen has a document. It is whether the document corresponds to daily practice at 02.00, 06.00, 09.00, and the moment a child opens the tray.

The partner rebellion did not become the main event

Our earlier piece, “Deadline Day: What Happened to MBG’s August 17 Kitchen Rebellion?”, found no verified evidence that the threatened national lockout began at scale. The new public record remains consistent with that reading.

The July ultimatum mattered because it exposed the relationship between BGN and its partner network. The partners were not only objecting to paperwork. They were objecting to unilateral policy, insentive structures, beneficiary counts, and the economic model of kitchen operation. That is why the 30-day evaluation cannot be assessed only as a sanitation audit. It also had to answer whether BGN could make the partnership governable.

On that narrow point, BGN seems to have succeeded in avoiding the visible August 17 rupture. There is no strong public evidence that a national kitchen lockout occurred. Service disruption instead came through government suspensions, sanitation closures, poisoning-linked closures, and later administrative rationalization.

But avoiding the lockout is not the same as resolving the conflict. A lockout is an acute symptom. The underlying pressure remains if SPPG incentives, beneficiary loads, supplier obligations, virtual-account controls, and partner approvals remain opaque.

The corruption nodes remain open

The 30-day evaluation did not publicly close the corruption questions.

The most important signal after the window is that enforcement outside BGN continued. Suara reported on 9 September that Kejaksaan Agung examined six witnesses in a corruption investigation into MBG governance at BGN for 2025–2026, including a seller of ompreng. The report said the case had seven named suspects and included alleged schemes involving jual beli titik SPPG and procurement of goods worth trillions of rupiah (Suara, 9 September 2026).

That sits directly on top of MBG Watch’s earlier “Ompreng Toll,” “One Failure, Not Four,” and “The Seven Suspects” analyses. The kitchen evaluation may have identified hygiene failures. It did not publicly show that BGN broke the procurement channels, forced-purchase dynamics, or approval-market incentives that those pieces tracked.

The KPK track is similar. CNN Indonesia reported that BGN and KPK met on 7 July to discuss BGN’s action plan for 10 recommendations from KPK’s MBG governance review. KPK’s concerns included the absence of a comprehensive blueprint, conflicts of interest, non-transparent recruitment, and an MBG ecosystem that had not been systematically built (CNN Indonesia, 7 July 2026). Warta Ekonomi’s report of the same process said KPK concerns included conflicts of interest in SPPG partner determination, weak transparency in partner verification and validation, weak food-safety supervision because of limited involvement by health offices and BPOM, and the absence of measurable short- and long-term success indicators (Warta Ekonomi, 7 July 2026).

Antara’s April report on KPK’s earlier governance findings is even more direct: KPK warned that MBG’s large budget was not matched by adequate regulation, governance, or supervision, creating risks of accountability failure, conflicts of interest, inefficiency, and corruption. KPK recommended comprehensive binding regulation, review of the bantuan pemerintah mechanism, stronger food-safety oversight involving health offices and BPOM, standard financial reporting, and measurable success indicators with baseline data (Antara, 17 April 2026).

The post-evaluation record does not show that those structural reforms were completed. It shows that BGN met KPK, formed an action-plan process, and took sanitation enforcement actions. It does not show that conflicts of interest, opaque approvals, procurement rents, or budget-execution controls were resolved.

This is the central governance finding: the evaluation was strongest where a kitchen could be inspected. It was weakest where a network had to be redesigned.

The beneficiary number is still not validated in public

One of the most important admissions before the evaluation was BGN’s own question: is the 63 million beneficiary number real?

On 21 July, iNews reported BGN Deputy Head Agustina Arumsari saying the evaluation would focus on validating the 63 million beneficiary figure and restructuring SPPG incentives. Her phrasing was unusually plain: “63 juta itu beneran nggak sih?” — is 63 million actually real? She said BGN needed to know the correct number, who should receive meals, which areas should be prioritized, and how SPPG incentives should differ by capacity and territory (iNews, 21 July 2026).

By 27 August, RRI reported one change: future beneficiary targets would be determined by BGN, not by each SPPG, and the 3B target group — pregnant women, breastfeeding mothers, and toddlers — was listed at 21,985,314 based on Health Ministry data (RRI, 27 August 2026).

That is progress in governance logic. Beneficiary counts should not be self-declared by kitchens that also depend on beneficiary volume for operating economics. But it is not yet beneficiary validation. The public still lacks a reconciled count by school, pesantren, posyandu, district, vulnerability category, delivery route, and funding allocation.

This matters because beneficiary fiction is not an administrative nuisance. It changes almost every denominator:

• cost per beneficiary;
• kitchen load;
• food-safety exposure;
• claims about coverage;
• procurement volume;
• staffing needs;
• whether a poisoning rate is rising or falling;
• whether MBG is reaching the children and mothers most in need.

A program can close hundreds of kitchens and still not know whether its core service population is real.

Budget execution and the September shock

The strongest post-window budget signal came on 17 September. Katadata reported BGN could temporarily close 12,512 SPPG for administrative and quality problems, and that MBG beneficiaries had fallen nearly 40 percent, from 51.85 million the previous month to 31.52 million. Reported reasons included SLHS status, multiple virtual accounts used to access MBG funds, beneficiary counts below 1,000, and disputes between foundation owners and SPPG operators (Katadata, 17 September 2026).

This is one of the most important facts in the whole record. It suggests BGN’s problem was no longer only unsafe kitchens. It was a basic operating-governance problem: accounts, operators, beneficiary thresholds, disputes, certification, and activation status.

A 40 percent reported beneficiary drop can mean two very different things.

It could mean BGN finally removed inflated or unsupported service claims from the ledger. If so, the program became more honest, even if the headline looked worse.

Or it could mean real beneficiaries lost service because BGN had allowed an unsafe and poorly documented kitchen network to grow too large before the controls were ready. If so, the cost of late governance landed on children and households rather than on the institutions that scaled the system too quickly.

The public record does not yet let us separate those two. That is why BGN needs to publish not only the number of kitchens closed, but the beneficiary disposition attached to each closure: transferred, paused, unverified, duplicate, ineligible, or still served by another kitchen.

Domain assessment under the Continuous Monitoring Framework

Nutrition outcomes

Not assessable from the evaluation record. The disclosures focus on kitchens, certificates, heads of unit, and service counts. They do not show measured nutrition outcomes, anthropometric baselines, dietary adequacy, or learning and health effects. MBG still lacks the outcome ledger KPK called for: measurable indicators and baseline data.

Food safety

Partially improved in enforcement; not yet improved in outcome. SLHS is now visibly enforced, thousands of kitchens have faced closure or suspension pressure, and BGN is naming SOP violations. But poisoning incidents continued at scale after the evaluation window. The system still needs route-time, temperature, storage, recall, and daily compliance proof — not just certification status.

Budget execution

Still structurally exposed. The September 17 report of multiple virtual accounts, disputed SPPG ownership, beneficiary thresholds, and possible closure of 12,512 SPPG shows that payment and operating controls remain central vulnerabilities. The Kejagung investigation into SPPG-point sales and ompreng-linked procurement shows that corruption risk remains active, not historical.

Equity of access

Unclear and potentially harmed. BGN says it will prioritize 3T activation and 3B groups, and that it will determine beneficiary targets centrally. But service contractions, kitchen closures, and the reported fall from 51.85 million to 31.52 million beneficiaries make equity impossible to assess without a district-level beneficiary transfer ledger.

Governance integrity

Improved in acknowledgment; incomplete in repair. BGN engaged KPK, formed an action-plan process, and acted against some non-compliant SPPG. It has not publicly demonstrated full implementation of the KPK governance agenda, partner-selection transparency, procurement repair, conflict-of-interest controls, or a public final evaluation dataset.

What the evaluation fixed

It fixed three things enough to count.

First, it made SLHS a live enforcement threshold. Before July, hygiene certification was too easy to treat as paperwork. After the evaluation, BGN and coordinating ministers were willing to attach closures to it.

Second, it made SPPG profiling harder to ignore. The 463 kitchens at risk for not submitting profiling show BGN trying to pull kitchen identity, beneficiary assignment, and operator accountability into one administrative frame.

Third, it broke the appearance that every kitchen was politically untouchable. Hundreds were permanently closed, nearly 2,000 were later temporarily closed, and hundreds of heads were removed. That matters in a program where local networks, foundations, operators, vendors, and officials all have incentives to keep volume moving.

What it did not fix

It did not produce a public final report.

It did not validate the 63 million beneficiary number in public.

It did not show a complete kitchen-by-kitchen status ledger.

It did not stop poisoning incidents after the evaluation window.

It did not publicly resolve cold-chain and route-time controls.

It did not close the ompreng and procurement questions.

It did not show implementation of the full KPK governance agenda.

It did not show whether the August 17 partner threat was resolved by reform, delayed by negotiation, or overtaken by BGN’s own suspension process.

The least-harm next step

BGN does not need another broad slogan about evaluation. It needs to publish the evaluation ledger.

The minimum public ledger should include:

1. every SPPG identifier, operator, location, and ownership/foundation relationship;
2. SLHS status, inspection date, failed criteria, correction deadline, and reopening condition;
3. beneficiary count before and after evaluation, with reason codes for any reduction;
4. incident history, including poisoning, late delivery, unsafe temperature, and complaint records;
5. funding account status, including duplicate or abnormal virtual-account flags;
6. procurement and equipment supplier relationships where BGN funds or mandatory standards create vendor dependence;
7. KPK recommendation status: not started, in progress, completed, independently verified;
8. district-level continuity record showing which beneficiaries were transferred, paused, or removed.

This should not be a PDF summary. It should be a machine-readable public dataset, with privacy protection for children and households, but full accountability for kitchens, operators, and state decisions.

If BGN can publish that, the 30-day evaluation becomes the beginning of structural reform. If it cannot, the evaluation remains what the current record suggests: a necessary sanitation intervention that bought time, reduced some visible risks, and left the deeper machine mostly unproven.

What I am uncertain about

The largest uncertainty is whether BGN has a complete internal evaluation report that has not been made public. The public record I reviewed shows significant announcements after the 30-day period, but not a final report with kitchen-by-kitchen outcomes or full governance implementation. If such a report exists, the assessment should change around what BGN fixed internally. Until it is disclosed, the accountability finding has to rest on what citizens, beneficiaries, schools, partners, and oversight bodies can actually inspect.