the accident reports that never come

Problem statement

Aviation's safety record is built on a learning loop: every accident is investigated by an independent state authority and a final report, published for the whole industry, converts one crash into prevention everywhere. Annex 13 to the Chicago Convention obliges the investigating state to file a preliminary report within 30 days and to make the final report public "as soon as possible and, if possible, within twelve months," with interim statements on each anniversary if it cannot. The loop is broken for about half of accidents. IATA's data show that of 268 accidents in its Annual Safety Report for 2018–2023 only 140 investigations (52 percent) were completed, and Flight Safety Foundation data for 2018–2022 show that of 448 accidents and serious incidents involving aircraft over 5,700 kg that required an Annex 13 report, only 160 (36 percent) were completed; in the ICAO Asia-Pacific region final reports were unavailable for 60 of 140 accidents (42.5 percent) over 2017–2023. The unsolved problem is institutional: every ICAO member state has agreed to the standard, ICAO can audit but not enforce it, and the states least likely to publish are those with the least investigative capacity and the strongest domestic reasons not to.

Why this matters

An unpublished report is a safety lesson withheld from every other operator, manufacturer, regulator, and air-navigation provider that flies the same aircraft in the same conditions; the industry papers put it directly — failure to publish "prevents operators, equipment manufacturers, regulators, infrastructure providers, and other concerned stakeholders from accessing critical information that could make aviation even safer." It also corrupts the data on which risk management runs: in the Flight Safety Foundation's Aviation Safety Network data for Asia-Pacific, "unknown" — the category assigned when no final report exists — "ranks as the 5th highest occurrence category in the region." And for the families of the dead it means never learning what happened — the victims' federation brought the issue to the 2025 ICAO Assembly on exactly that ground (industry advocacy material for the period counts 12 unpublished accidents that killed 242 people; see Source Notes).

What’s been tried and why it hasn’t worked

The obligation has existed for decades and its non-fulfilment has been raised at every level: IATA has made Annex 13 compliance a strategic priority and pressed it through its Director General, its AGM, regional safety groups, and the ICAO Air Navigation Commission, and has developed a five-point roadmap with ICAO's Accident Investigation Panel "to understand why accident reports are not published"; in February 2025 seven industry bodies (ACI Europe, ECA, FSF, IATA, ICCAIA, IFALPA, IFATCA) issued a joint position paper; ICAO regional accident-investigation groups have been urging compliance and building training; and the victims' federation took the issue to the 2025 Assembly asking that ICAO strengthen monitoring through its Universal Safety Oversight Audit Programme and "identify the reasons why final reports are not being published." Progress is marginal — FSF's Asia-Pacific series shows only "a slight improvement over the analysis period from 2017 to 2023." The reasons the papers name are structural: many states have not established the independent investigation authority Annex 13 requires; qualified investigators are scarce (in the ICAO APAC region the USOAP effective-implementation score is 47.6 percent overall and 53.61 percent for the "qualified technical personnel" element); some states let a preliminary report stand in for a final one; and the treaty gives ICAO no sanction beyond audit findings and moral suasion, so a state facing a report that would embarrass its airline, regulator, or courts pays no price for silence. Exhortation, the only tool so far applied, does not change those incentives.

What would unlock progress

Progress requires changing the incentives and the capacity together: a transparent, continuously updated public ledger of overdue reports by state (so non-publication has reputational cost in a system that runs on reputation), a pooled or regional investigation capacity that small states can draw on so that "we lack investigators" stops being the reason (IFALPA offers 178 accredited pilot-investigators; regional accident-investigation organizations exist in some regions), and a systematic diagnosis — which the papers explicitly say has not been done — of why each missing report is missing (capacity, judicial interference, political sensitivity, loss of evidence). Adjacent precedents worth examining: maritime casualty-investigation reporting under IMO, and the publication of ICAO's own USOAP audit results, which turned a confidential audit into a reputational instrument.

Entry points for student teams

A data team could build the missing dataset — merge Aviation Safety Network, IATA, and ICAO records to produce a per-state, per-year ledger of Annex 13 reports due, published, late, and never published, and model what predicts non-publication (state income, USOAP score, existence of an independent authority, whether the operator or manufacturer is domestic, whether criminal proceedings were opened) — as a research contribution that the industry papers say does not yet exist. A policy-design team could draft a proposal for a regional pooled investigation service for a specific region (e.g., Pacific island states) with governance, funding, and Annex 13 delegation mechanics. A design/communication team could prototype the public "overdue reports" dashboard and test whether it changes behavior in a simulated setting. Relevant skills: aviation safety, data science, international law and institutions, public policy.