Who Pays for the Shelter Meal? The Reimbursement Ledger MBG Needs

MBG Watch · 2026-08-17

The premise

On 16 August 2026, after the Magnitude 7.7 earthquake centered near Nagekeo, public reports said the Indonesian government had identified 71 Satuan Pelayanan Pemenuhan Gizi (SPPG) or MBG kitchens that could help feed evacuees in NTT. RRI quoted Home Affairs Minister Tito Karnavian saying some schools were closed or damaged, that he had communicated with BGN head Sudaryono, and that “71 SPPG” might be used; meals that would normally go to schools could be redirected to evacuation sites. VIVA carried the same basic account and added that BGN told affected SPPGs to put safety first, report conditions, coordinate with BNPB and local governments, and expect revisions to BGN Circular Letter No. 7 of 2025 on MBG service during natural-disaster emergencies.

There is also local evidence that this was not only a national coordination idea. Kompas reported that SPPG kitchens in Sikka distributed ready-to-eat meals to residents sheltering after the earthquake at Gelora Samador Maumere on 15 August, with four SPPGs expected to prepare about 300 portions each, or about 1,200 portions, and with Sikka considering the involvement of 18 operating SPPGs.

That is a useful public function. MBG Watch’s concern is not that kitchens helped. The concern is that a school-meal kitchen becomes a disaster-feeding kitchen the moment it buys extra rice, uses fuel, shifts staff time, serves an evacuation point, or records a child’s ordinary meal as a relief meal. At that point, the question is no longer only operational. It is also budget execution: who pays, who gets reimbursed, and how the public can tell the difference between MBG reach and emergency relief.

What the evidence supports

The public record supports four narrow claims.

First, government actors publicly considered temporary use of MBG/SPPG capacity for evacuee feeding in NTT. RRI and VIVA both reported the 71-SPPG figure and the ministerial logic: closed or damaged schools reduced ordinary school distribution, while evacuees had immediate food needs. That does not prove all 71 kitchens served meals. It does show that SPPGs were being treated as part of the emergency-feeding option set.

Second, at least some SPPG food support appears to have occurred locally in Sikka. Kompas reported prepared meals distributed to evacuees, named Gelora Samador Maumere, quoted Sikka SPPG coordination, and gave the 1,200-portion estimate. This is enough to justify a ledger question. It is not enough to infer total emergency-meal volume across Flores.

Third, BGN itself has previously described MBG kitchens as emergency-useful infrastructure. In February 2026, BGN published an article on SPPG Palembayan Salareh Aia in Agam, describing how a kitchen that later became an SPPG had already been used as a disaster kitchen after the November 2025 landslide and flash flood in Sumatra and Aceh, including support with LPG cylinders, difficult routes, and a generator used by residents and schoolchildren to charge phones. The same article matters because it shows that “emergency kitchen” use is not a one-off improvisation. It is now a pattern BGN can see.

Fourth, the disaster-law basis for feeding evacuees is clear, even if the MBG reimbursement pathway is not. Law No. 24 of 2007 defines disaster emergency response as immediate activity that includes basic-needs fulfilment, protection, management of evacuees, and recovery of facilities. Its explanation of Article 53 lists basic needs as including clean water and sanitation, food, clothing, health services, psychosocial services, and shelter. BNPB Regulation No. 7 of 2008 is specifically titled as a procedure for assistance to fulfil basic needs. Feeding evacuees is therefore not a discretionary kindness outside the disaster system. It is part of disaster response.

What the evidence does not support

The evidence does not show that the Flores emergency use of SPPGs created a national fiscal shock, a currency shock, or a macroeconomic event. Rupiah Stability Watch’s adjacent analysis is useful because it names a channel: emergency feeding can become fiscally relevant if local obligations, supplier prices, fuel, water, cold-chain costs, and reimbursement delays scale without records. But the Flores evidence available on 17 August 2026 supports a narrower MBG conclusion: the accounting channel should be made inspectable before scale and duration make it harder to reconstruct.

The evidence also does not show that MBG caused harm in Flores, that emergency feeding was inappropriate, or that kitchens should avoid helping evacuees. In an emergency, unused safe kitchen capacity is valuable. The least-harm position is to let safe kitchens help while preventing emergency meals from disappearing into ordinary MBG numbers.

Nor does the public record answer the budget-line question. Public sources show coordination among BGN, the Ministry of Home Affairs, BNPB, and local governments. They do not show whether emergency meals supplied by MBG kitchens would be paid from the ordinary MBG operational allocation, local/provincial disaster funds, BNPB/BPBD support, social-assistance channels, a later BGN reimbursement, or a mixed pathway. That absence is the point of the ledger.

The reimbursement ledger BGN should publish

BGN does not need to expose children, evacuees, names, faces, ID numbers, or precise household locations to make emergency feeding accountable. The public record can be operational and financial without becoming surveillance. Building on MBG Watch’s earlier privacy piece, “Seen Without Being Watched,” the ledger should be aggregated at kitchen, batch, shelter or distribution-point level.

A minimum public reimbursement ledger should include:

1. Event and command-post reference: disaster event, emergency status period, responsible command post, and the BGN/BPBD contact responsible for reconciliation.

2. Kitchen identity and status: SPPG ID, district, whether the kitchen was directly affected, whether it was operating normally, partially operating, paused, or assigned to emergency support.

3. Distribution point: shelter or public distribution point name, not beneficiary names. Where safety requires it, use a coded location known to the command post.

4. Meal classification: ordinary MBG school meal, replacement meal for a temporarily closed school, evacuee relief meal, responder meal, or other category. This is the line that prevents double-counting.

5. Batch record: production date and time, batch ID, menu or substitution, estimated shelf-life window, dispatch time, and receiving-point acknowledgement.

6. Meal counts: portions planned, portions produced, portions delivered, portions unserved or discarded, and the reason for any material difference.

7. Unit cost: food ingredients, packaging, labor, logistics, and any exception costs such as fuel, LPG, water trucking, generator fuel, ice, cold-chain rental, route diversion, or temporary storage.

8. Supplier and source: whether ingredients came from ordinary MBG suppliers, emergency procurement, donations, redirected stock, or local government/BNPB support.

9. Payer and reimbursement status: initially charged party, expected payer, amount submitted, amount verified, amount paid, amount rejected, amount still unpaid, and days outstanding.

10. Safety and corrective actions: kitchen damage, water/power interruption, menu substitution because of supply disruption, temperature-control failure, stop-go decision, complaint, illness signal, or corrective action.

11. Ordinary-service protection: which schools or beneficiary groups did not receive ordinary MBG service because capacity was redirected, and what replacement arrangement was made.

This is not a demand for bureaucratic delay before feeding people. It is a record to be filled as the work happens and reconciled afterward. In practice, the first version can be a daily CSV or dashboard: one row per kitchen, per batch, per destination, per payer status.

Why the ledger matters for budget execution

BGN’s own January 2026 statement said MBG funds were being disbursed directly from the state treasury to SPPGs, with 70 percent used for raw materials, 20 percent for operations including volunteer wages, and 10 percent for partner incentives. That structure may work for ordinary school-feeding if meal counts, supply plans, and beneficiary lists are stable. Disaster feeding changes the cost object.

An evacuee meal may use the same kitchen, people, and suppliers, but it is not the same accounting event as an ordinary school meal. It may require different delivery points, additional packaging, new water arrangements, more fuel, substitute ingredients, longer holding time, generator use, damaged routes, or payment to suppliers outside the normal rhythm. If those costs remain inside ordinary MBG expenditure, the public cannot tell whether MBG expanded service, shifted service, depleted ordinary school-meal funds, or created an unpaid claim against some later disaster budget.

The risk is not only central-government accounting. It also lands on the kitchens and vendors. If an SPPG partner is asked to serve evacuees first and reconcile later, the safest and most willing kitchen may carry the working-capital burden. Small suppliers may wait for payment, extend informal credit, or raise prices to compensate for uncertainty. Local governments may book the benefit without booking the obligation. BGN may count output without showing whether it was ordinary nutrition-program output or emergency-relief output.

That is how a good act becomes an opaque ledger. The solution is not to stop the act. It is to classify it.

The least-harm path

For the current Flores response, BGN and local command posts can take a proportional path.

First, publish a short emergency-feeding notice for each affected district: which SPPGs are assigned, which are paused for safety, which are serving schools, and which are serving shelters.

Second, separate meal counts from the first day. Ordinary school meals, school-replacement meals, and evacuee relief meals should never sit in one undifferentiated “MBG portions served” number.

Third, publish reimbursement status weekly until all emergency claims are resolved. A kitchen that served evacuees should be able to see whether the payer is BGN, BPBD/APBD disaster funding, BNPB support, social assistance, or another channel. If the channel is mixed, say so row by row.

Fourth, protect privacy by design. Publish shelter-level or coded distribution-point totals, not beneficiary identities. Keep any personally identifiable data inside the authorized disaster-response system, subject to need-to-know access and retention limits.

Fifth, do not penalize safe kitchens for helping. If a kitchen can document emergency service, safety controls, and actual costs, reimbursement should be predictable. If a kitchen cannot document basic batch, meal, and cost records, the correction should be training and tighter controls before expansion, not retrospective blame without standards.

Sixth, revise BGN’s emergency-service circular in public. VIVA and RRI reported that BGN expected to revise Circular Letter No. 7 of 2025. The revision should not only tell kitchens to coordinate with BNPB and local governments. It should also define the reimbursement ledger, the cost categories, the privacy boundary, and the rule for separating ordinary MBG meals from disaster-relief meals.

What I am uncertain about

The most important uncertainty is the payment pathway. I found public evidence of SPPGs being prepared for and used in emergency feeding, and public evidence that disaster response includes basic-needs fulfilment such as food. I did not find a public BGN or government ledger stating which budget line will reimburse SPPGs for Flores emergency meals.

The second uncertainty is scale. The 71-SPPG figure is a preparedness or potential-use figure reported from coordination statements. The Kompas Sikka report gives a concrete local meal estimate. The public record available here does not establish total SPPG emergency meals across NTT.

The third uncertainty is whether ordinary school-meal service was materially displaced. Public reports say some schools were closed or damaged, and that meals could be redirected to evacuees. They do not give a district-by-district account of which school beneficiaries missed, received, or had replacement meals.

Those uncertainties are not reasons to distrust the emergency response. They are reasons to publish the ledger while memories, batch sheets, supplier invoices, and command-post decisions are still fresh.

The standard is simple: when an MBG kitchen feeds a shelter, the meal should be counted once, paid once, and visible enough to protect children, evacuees, kitchens, vendors, and public money.