When Disaster Breaks the Beneficiary List: The Identification Boundary MBG Needs in Emergency Feeding

MBG Watch · 2026-08-30

The premise

A disaster does not only move people. It moves the evidence.

The current Nepal–Tibet flood and mudslide reporting is useful for MBG Watch only as a systems comparator, not as a claim about Indonesia. As of 30 August 2026, The Guardian reported a combined toll of 691 dead and more than 3,000 missing in Nepal and Tibet, with poor weather hampering aerial rescue and with specialized tunnel and rescue capacity entering the response. Al Jazeera’s earlier 27 August explainer reported lower but already severe figures: more than 390 killed and more than 1,500 missing, with helicopters, drones, search dogs, damaged roads, high river levels, and deep sediment complicating rescue.

The narrow lesson is not that MBG has failed in a comparable way. It is that when a disaster breaks ordinary registers, public systems need to know which record is doing which job. A meal-delivery record is not a missing-person record. A beneficiary-validation record is not a rescue-intelligence record. A reimbursement record is not a health-surveillance file. If those boundaries blur, the public may get neither accountability nor privacy.

Indonesia already has disaster data machinery that recognizes this distinction. BNPB’s 2023 implementation guide for disaggregated displacement data says the disaster command post is authorized to provide data and information on emergency response, including disaggregated displacement data, and defines that data as information separated by sex, age group, and protection needs. The same guide also treats child identity data, government identification numbers, health information, and geolocation as sensitive personal data. That is the line MBG should preserve if its kitchens feed shelters, school replacements, or displaced non-school beneficiaries.

What the public record already shows

BGN has publicly described SPPG kitchens entering disaster feeding in at least four ways.

First, BGN reported that an SPPG in North Cikarang delivered MBG to an evacuation post for about 150 flood-affected residents, while also continuing to serve regular program beneficiaries. The report says meals moved from the SPPG kitchen to tents with help from Babinsa and Bhabinkamtibmas, and frames SPPGs as front-line nutrition distribution when household food production is disrupted.

Second, BGN’s Agam reporting shows the boundary between ordinary school service and post-disaster school replacement. One article says MBG was being distributed regularly to students in an emergency school after flood damage. Another says a kitchen that later became an SPPG had already been used as an emergency disaster kitchen, and that the first Ramadan MBG distribution reached students whose school was operating from a mosque after the old building was swept away.

Third, BGN reported that more than 725,000 MBG portions were distributed to disaster victims in North Sumatra and Aceh, with 53 of 133 SPPGs in North Sumatra and 91 of 314 SPPGs in Aceh temporarily redirected to ready-to-eat food for flood and landslide victims. The statement says SPPGs were prepared not only for routine service but also for extraordinary conditions, and would work with local governments, disaster-management institutions, and other parties.

Fourth, BGN’s 2026 technical guidance, as indexed and downloaded from BGN’s own CDN, describes SPPGs as needing to be adaptive in special or emergency conditions, including natural disasters, and ready to be at the front line for MBG service to affected communities, victims, or evacuees according to established direction and coordination. The same guidance also defines normal beneficiary groups, including school-age students, pesantren students, children under five, pregnant women, breastfeeding mothers, and other targeted groups.

These records are enough to show that disaster feeding is no longer hypothetical for MBG. They are not enough to show that BGN has published a public ledger that distinguishes five different categories: ordinary school MBG meals; school-replacement meals when a school is closed, moved, or damaged; shelter meals; non-school care-route meals for pregnant women, breastfeeding mothers, toddlers, or other eligible groups; and meals served when identity or household status is temporarily uncertain because of evacuation.

That absence matters because BGN is also moving toward stronger beneficiary identity validation. Its 13 August 2026 statement with the Ministry of Home Affairs says the program is integrating Dukcapil data to improve targeting, prevent duplicate beneficiary records, and move beyond counting “how many” recipients toward knowing “who” receives benefits under a “one person, one beneficiary” principle. That may be reasonable in ordinary administration. In a disaster, it needs a privacy-preserving emergency mode.

The existing Indonesian boundary MBG can borrow

BNPB and health-sector practice already point to a safer public shape.

BNPB’s 2008 basic-needs assistance rule defines vulnerable groups to include infants, children under five, children, pregnant or breastfeeding mothers, persons with disabilities, sick people, and older people. It says basic-needs assistance should prioritize vulnerable groups; food assistance may be provided as foodstuffs or cooked meals from public kitchens; cooked food from public kitchens is two meals per day; and the food value is equivalent to 2,100 kilocalories.

BNPB’s 2012 logistics-management rule asks emergency logisticians to identify deaths, missing persons, injuries, and displacement separately. For displaced people, it calls for data by sex, age, and vulnerable group, including toddlers, children including those with special needs, pregnant women, older people, and people with disabilities. That structure is important: missing persons and evacuee needs are related in operations, but they are not the same record.

The Ministry of Health’s crisis center reports show the same pattern in health surveillance. Its Semeru update counted evacuees, health complaints such as acute respiratory infections and gastritis, and vulnerable groups including pregnant women, babies, toddlers, and older people. Its Cianjur earthquake update counted 79,352 evacuees across 340 sites and said the government classified evacuees by age and vulnerability, while recording health personnel including surveillance staff.

Dukcapil practice adds the identity-restoration side. In West Sumatra, Ditjen Dukcapil reported issuing 5,795 civil-registration documents after floods, including family cards, child identity cards, birth certificates, death certificates, digital identity activations, electronic ID cards, and new biometric recordings. A separate Dukcapil statement asked affected provinces to map which local Dukcapil offices could still serve, identify damaged facilities, report document needs to replace lost or damaged identity papers, and provide periodic reports.

Taken together, these records suggest a usable boundary. Disaster response may need identifiable records to restore legal identity, reunify families, protect children, classify health needs, and deliver aid. But the public MBG ledger does not need to expose those identities. It needs to show whether food service remained valid when the ordinary school or household reference broke.

The least-harm ledger

The public ledger MBG needs in emergency feeding should be narrower than the operational record and stronger than a press release.

It should publish, at SPPG and response-area level, five separated lines.

1. Operating status: whether the SPPG is in normal service, school-replacement service, shelter-feeding service, non-school care-route service, mixed service, suspended service, or recovery service. This extends MBG Watch’s earlier Flores pieces on shelter kitchens, reimbursement, and aftershock operating status.

2. Meal route and eligible group: counts by route and group, not by name. The route would distinguish ordinary school delivery, relocated school delivery, shelter delivery, mobile/community delivery, and referral delivery through health or social-service posts. The group would distinguish students, toddlers, pregnant women, breastfeeding mothers, and other authorized beneficiaries without publishing individual identity.

3. Validation basis: whether counts are based on the ordinary school roll, a temporary school register, shelter aggregate data, health/social-service aggregate referral, Dukcapil-restored identity verification, or emergency provisional counting pending correction. This builds on “Seen Without Being Watched” and “Before the Number Becomes a Fact”: the public needs provenance, not exposure.

4. Safety and suitability status: whether the meals met the same food-safety, nutrition, transport-time, menu, and special-diet controls expected in ordinary service, and whether the kitchen had to use emergency adaptations. This should include correction entries when a count or safety status changes.

5. Funding and reimbursement line: whether the meal is charged to ordinary MBG service, emergency relief, school-replacement continuity, local-government coordination, or a later reimbursement process. This extends “Who Pays for the Shelter Meal?” and also matters for Rupiah Stability Watch’s crossing: emergency feeding becomes currency-relevant infrastructure only when fiscal and logistics ledgers remain legible under stress.

The private operational record can be more detailed. It may need names, NIK checks, family-card recovery, medical vulnerability, pregnancy or breastfeeding status, missing-person reports, child-protection referrals, and exact shelter or household location. But those fields should sit behind the proper disaster, health, civil-registration, and child-protection authorities. They should not leak into a public MBG transparency interface.

A simple rule would help: publish the meal proof, not the person. Publish the route, count, status, source basis, correction trail, and funding line. Do not publish missing-person lists, child identities, pregnancy or health status, NIK or family-card details, precise household location, rescue-sensitive movement, or shelter-level details where a small count could identify a child or family.

What this would prevent

A boundary like this prevents three quiet failures.

The first is false certainty. If an evacuated child receives a meal in a shelter while the school register is unusable, the public number should not pretend to be an ordinary school-distribution number. It should be marked as emergency provisional or shelter-aggregate until corrected.

The second is double counting. A child may appear on a school roll, a shelter list, a temporary school list, and a household relief record. BGN’s Dukcapil integration is explicitly meant to prevent duplicate beneficiary records. In disaster mode, the public ledger should show the validation basis and correction mechanism without revealing the child.

The third is privacy drift. Once a feeding program starts borrowing missing-person, health, pregnancy, or child-protection data, it can become tempting to prove success by exposing too much. The BNPB displacement-data guide already warns that child identity, official identification numbers, health information, and geolocation are sensitive. MBG should treat that as a design constraint, not an afterthought.

What the evidence does not support

The evidence does not support saying that Nepal or Tibet predicts an Indonesian MBG failure. It does not support saying that BGN lacks operational coordination in every disaster setting. BGN has publicly reported disaster feeding, temporary SPPG redirection, continued regular service beside emergency service, and coordination with local governments and disaster agencies.

The evidence also does not support treating identity validation as inherently harmful. Restoring identity after disaster can protect rights. Dukcapil’s post-disaster document work is evidence of that. The issue is not whether government may validate identity. The issue is which parts of that validation become public, which stay operational, and how uncertain counts are corrected without exposing people.

What I am uncertain about

I could not verify, from public BGN pages searched and retrieved for this piece, a disaster-identification circular or public ledger schema that separates ordinary MBG meals, school-replacement meals, shelter meals, non-school care-route meals, and emergency provisional meals where identity is uncertain.

I could verify that BGN’s technical guidance recognizes emergency and disaster adaptability for SPPGs, and that BGN has public articles describing disaster feeding. I could not verify whether BGN has an internal SOP, not public, that already draws the identification boundary described here.

I also could not verify the exact current Nepal–Tibet toll from a primary disaster-authority source. For that reason, this piece uses The Guardian and Al Jazeera only to establish the comparator signal: a fast-moving disaster with high death and missing counts, damaged access, rescue complexity, and identification pressure.

The practical test

The next useful public test is modest. When BGN next reports disaster feeding, the report should allow a reader to answer five questions without learning any displaced child’s identity.

Was this an ordinary MBG meal, a school-replacement meal, a shelter meal, a non-school care-route meal, or an emergency provisional meal? Which eligible group was served? What validation basis supports the count? What safety and correction status applies? Which budget or reimbursement line carries the cost?

If those five answers are visible, the public can see whether the meal count remains trustworthy when the beneficiary list breaks. If the identities stay protected, the child can be seen without being watched.