The Canteen Pivot: Measuring BGN’s Post-Corruption Operational Change Against the Evidence

MBG Watch · 2026-10-02

The premise

On 15 July 2026, Badan Gizi Nasional said it was studying “various alternatives” for the implementation of Makan Bergizi Gratis, including the possible involvement of school canteens. The agency framed the review as part of an evaluation of effectiveness, efficiency, reach, food-safety standards, nutrition quality, and governance. It also said any change would be gradual, evidence-based, and cautious.

That language matters. After the corruption findings around MBG’s kitchen and procurement model, the canteen proposal should not be read as a small delivery tweak. It is an operational pivot being considered at the same time BGN is trying to answer KPK’s governance findings, the food-safety record, and the equity problem that has shadowed the program from the start.

The right question is not whether school canteens are good or bad. Some are clean, close to students, already known by school communities, and cheaper to mobilize than a newly built SPPG kitchen. Others are informal vendors with weak hygiene systems, limited equipment, and no capacity to produce standardized meals at MBG scale.

The question is narrower and more useful: does the canteen model correct the failure pattern already visible in MBG, or does it move the same risks into a less visible venue?

My reading is this: a canteen model could be a real reform only if BGN uses it to replace a centralized licensing-and-procurement bottleneck with a transparent, locally verified, safety-certified operating model. If it becomes a fast way to count more service points without publishing eligibility, inspection, payment, and incident data, it will reproduce the same problem in smaller units.

What the proposal actually says

The official BGN release does not announce implementation. It announces a study. Agustina Arumsari, BGN’s deputy head, said BGN was reviewing options so MBG could become more effective, efficient, and able to reach beneficiaries while maintaining food-safety standards, nutrition quality, and good governance. On canteens specifically, BGN said the review would cover infrastructure readiness, management capacity, hygiene and sanitation standards, and oversight mechanisms.

Those four tests are the correct ones. They are also the difficult ones.

A school canteen is not automatically a public feeding kitchen. It is usually designed for retail food access inside a school, not for guaranteed production of complete meals under a national entitlement. Moving MBG into canteens changes the control surface:

• from a dedicated SPPG kitchen to a mixed school-food environment;
• from centralized kitchen approval to many school-level operating decisions;
• from one food-production chain to multiple combinations of school staff, vendors, suppliers, and local oversight;
• from a relatively visible MBG facility to an existing school space whose records may be thin.

That can be good. It can also be dangerous. The difference is whether BGN treats canteens as infrastructure to be audited, not merely as capacity to be used.

Equity: canteens may help some 3T areas, but they are not an equity strategy by themselves

The strongest case for involving canteens is proximity. A canteen sits where children already are. If a school has a functioning kitchen or food-service space, using it could shorten the route between cooking and consumption, reduce transport time, and make parent-school oversight easier. In remote or island areas, that may be better than depending on a distant SPPG kitchen whose delivery route is fragile.

But the equity risk is just as clear. Canteen capacity is not evenly distributed. UNICEF’s school nutrition environment assessment found significant gaps in school food environments in Indonesia: few schools offered healthy food options, ultra-processed products were widely sold and promoted, and many schools lacked the facilities or trained staff needed to prepare meals on-site. A model that depends on existing school food infrastructure will therefore tend to help schools that already have infrastructure first.

That is the central equity test. If BGN measures success by the number of canteens onboarded, the model will look effective in urban and peri-urban schools before it has solved the harder access problem. Children in schools without canteens, without clean water, without adequate storage, or without trained food handlers could become administratively invisible: not excluded by policy, but delayed by design.

A credible canteen pivot therefore needs a negative map, not only a partner map. BGN should publish not just which schools can participate, but which schools cannot yet participate and why:

• no canteen or kitchen space;
• insufficient water, sanitation, or drainage;
• no safe storage or cold chain;
• no trained food handler;
• distance from certified supply routes;
• special-diet or disability access gaps;
• school status outside the ordinary administrative route.

Without that exclusion map, the pivot can widen the gap while appearing to expand the program.

Food safety: the failure was never only the kitchen type

The food-safety record makes the canteen proposal more serious than an efficiency measure. A May 2026 report citing Kementerian Kesehatan recorded 37,673 suspected food-poisoning victims from MBG through 10 May 2026, across 445 suspected incidents in 210 districts and cities in 36 provinces. It also reported that only 56.72 percent of operating SPPG units had a Sertifikat Laik Higiene Sanitasi.

KPK’s own regulatory analysis had already identified the same structural weakness earlier in the program’s life. It argued that food safety depends on daily kitchen practice across procurement, storage, preparation, cooking, and distribution; training alone is not enough. It also noted that local health-office oversight and SLHS certification had not been used strongly enough to assess kitchen feasibility and prevent poisoning.

This is the key point for canteens: the risk is not solved by changing the room where cooking happens. It is solved by changing the control system around that room.

A canteen model may reduce some risks if it shortens delivery time and uses already functioning school facilities. It may increase risks if canteens are admitted without the same or higher standards for:

• hygiene and sanitation certification;
• water quality;
• food-handler training and medical fitness;
• supplier traceability;
• temperature control;
• allergen and special-diet handling;
• daily batch records;
• incident reporting;
• independent closure authority.

There is also a governance detail that should not be missed. If the same agency that is under pressure to expand coverage also certifies canteen readiness, approves payment, and investigates failures, the conflict remains. KPK’s analysis warned that BGN’s dual role in regulating and monitoring SPPG created ambiguity and a risk of downgraded standards when expansion targets dominate. That warning applies just as strongly to canteens.

The least-harm version of the pivot would give local health offices a formal gatekeeping role before any school canteen serves MBG meals. BGN can coordinate and fund. It should not be the only institution deciding whether a kitchen is safe.

Governance: this could answer KPK, but only if it reduces central bottlenecks

KPK’s July 2026 follow-up with BGN is important because it turns the canteen proposal into a governance test. RRI reported that BGN submitted an action plan to KPK on 7 July to follow up ten recommendations from a review received on 17 March 2026. KPK said it would supervise, accompany, and monitor implementation. BGN said some recommendations had begun to be acted on, including improving beneficiary-data validity and payment mechanisms to prevent budget leakage.

The canteen proposal touches several of the same governance issues.

First, KPK’s regulatory analysis criticized the lack of comprehensive regulation and the centralized pattern of implementation through BGN. It recommended that BGN act as coordinator rather than sole implementer, with stronger cross-ministry and local-government roles.

A canteen model could support that recommendation if it makes school feeding more locally governed: schools, local health offices, education offices, parents, and suppliers each carrying defined responsibilities. But it would fail the test if “canteen partner” becomes simply a new label inside the same centralized approval machine.

Second, KPK identified transparency and accountability risks in the role of partners, SPPG operations, funding, monitoring, and beneficiary data. Canteens do not remove those risks. They multiply the number of nodes where those risks can occur.

That means every participating canteen should have a public minimum record:

• school name and location;
• operator identity;
• certification status;
• inspection date and inspecting authority;
• daily meal volume;
• payment channel and amount;
• supplier list;
• incident history;
• corrective actions;
• current operating status.

This is not bureaucratic decoration. It is how the public knows whether the pivot is reform or displacement.

Third, procurement risk does not disappear when production shifts closer to schools. The corruption machine documented in the SPPG model depended on approvals, required purchases, equipment flows, and opacity. A canteen model could bypass some of that if it avoids forced capital packages and lets schools use existing safe infrastructure. But it could also create a new market for “approved” canteen upgrades, kitchen kits, certification intermediaries, and preferred suppliers.

The warning sign would be a standard package sold as readiness. The reform sign would be a published technical standard with multiple ways to comply, open inspection records, and no mandatory vendor bundle.

The conditions under which the canteen model would improve MBG

The canteen proposal deserves a fair test. It could improve MBG under five conditions.

First, it must be piloted, not rolled out by announcement. The pilot should include different school types: urban schools with mature canteens, peri-urban schools with partial facilities, rural schools with weak infrastructure, and 3T contexts where logistics are the point of the exercise. If the pilot only selects schools already likely to succeed, it will answer the wrong question.

Second, canteen participation must be conditional on independent safety clearance. The clearance should include SLHS or an equivalent public-health certification route, water and sanitation checks, food-handler training, and an inspection schedule. BGN should not be able to waive these because coverage targets are behind.

Third, the model must publish who is not reached. Equity cannot be inferred from aggregate meal counts. BGN should report canteen eligibility, rejection, and remediation by district and school type. The children left behind by the canteen model are the most important metric in the first year.

Fourth, the payment system must be direct, auditable, and separable from procurement. If canteens are paid through opaque intermediaries, the model will inherit the leakage risk KPK flagged. If payments are tied to verified meals, beneficiary records, and public reconciliation, the model can become cleaner than the old route.

Fifth, incident authority must be independent. A canteen with a poisoning incident or repeated non-compliance should be paused by a health authority, not only by the program operator. Parents and schools need a visible way to report unsafe food without losing access to the program itself.

The conditions under which it would repeat the old failure

The canteen pivot would be cosmetic if BGN does any of the following.

It would repeat the equity failure if it counts easy urban schools first and leaves infrastructure-poor schools for a later, unfunded phase.

It would repeat the food-safety failure if canteens are admitted before hygiene certification, water, storage, and handler controls are real.

It would repeat the corruption failure if participation requires bundled equipment, preferred vendors, or informal facilitation fees.

It would repeat the accountability failure if BGN keeps approval, monitoring, payment, and incident investigation inside one closed chain.

It would repeat the data failure if the public sees aggregate meal numbers but cannot see which schools are served, which are pending, which failed inspection, and which suffered incidents.

These are not objections to canteens. They are the minimum tests of whether canteens are a reform.

What to watch next

The next evidence should be specific. BGN’s July statement says the canteen option is still under study. The study should produce records that can be checked.

The first record is the eligibility standard. What exactly must a canteen have before it can serve MBG? If the answer is mostly narrative — “ready,” “adequate,” “coordinated” — the standard is not yet real.

The second record is the pilot list. Which districts and school types are included? A credible pilot contains hard cases, not only convenient ones.

The third record is the inspector. Who signs safety clearance: BGN, the health office, the school, or a third party? This answer will show whether BGN has accepted KPK’s point about distributed governance.

The fourth record is the payment trail. Are canteens paid directly? Are suppliers named? Are meal counts reconciled against beneficiaries? This is where the corruption machine either loses oxygen or changes shape.

The fifth record is the exclusion ledger. Which children remain outside the canteen model, and what funded fix is attached to each exclusion reason? This is where equity becomes measurable.

What I’m uncertain about

The public record I reviewed confirms that BGN is studying the canteen option and names the right review categories, but it does not yet show the operational design. That means the proposal cannot fairly be judged as success or failure. It can only be judged against the conditions it would need to meet.

I am also uncertain about the current national distribution of usable school canteens. The available school-food-environment evidence shows large gaps in facilities and trained staff, but it does not by itself provide a complete canteen-by-canteen readiness map. That map is exactly what BGN should publish before implementation.

Finally, the corruption risk depends on procurement design that has not been disclosed. If the canteen model avoids mandatory equipment bundles and uses existing safe infrastructure, it could weaken the extraction channels seen in the SPPG model. If it creates a new certification-and-upgrade market without public records, it could become another toll gate.

The bottom line

The canteen proposal is worth studying because it points in the right direction: closer to schools, potentially closer to local oversight, potentially less dependent on the compromised SPPG expansion machine.

But closeness is not governance. Existing infrastructure is not automatically safe infrastructure. A new venue is not a reform unless the authority, records, and incentives change with it.

The canteen pivot should be treated as a test of BGN’s post-corruption repair. If it publishes the eligibility map, safety clearances, payment trail, exclusion ledger, and incident record, it could become the first visible sign that MBG is moving from expansion-by-command to delivery-by-proof. If it does not, the same children will remain hardest to reach, the same safety controls will fail under pressure, and the same corruption risks will learn a new address.