Humanitarian Access

What this driver is

Humanitarian access is the continuing ability of civilians to receive, and of impartial relief actors to deliver, assistance and essential services in armed conflict, siege, disaster, and displacement. It is a mechanism, not a headline and not a welfare score. It is not a single convoy, not a donation appeal, and not “the humanitarian situation” as a mood. The question it asks is operational: can food, water, medicine, fuel, and medical evacuation move to the people who need them, with the consent or at least the non-obstruction of the parties who control the ground, the air, and the crossings.

The legal core is international humanitarian law, not a development-finance ratio. Fourth Geneva Convention rules on relief (including Articles 23 and 59) and Additional Protocol I Article 70 treat relief as a duty of parties toward civilians, subject to technical arrangements and to the prohibition on using starvation of civilians as a method of warfare. UN Security Council practice, including resolutions on famine and conflict (notably 2417) and successive Gaza and Syria access texts, treats obstruction, delay, and diversion as a governance problem with dates and crossing names. OCHA’s access reporting — movement requests granted or denied, checkpoints, visas, telecommunications blackouts, attacks on remaining health facilities — is the observational language of the driver.

This driver is not External health expenditure as a share of current health spending (World Bank SH.XPD.EHEX.CH.ZS). It measures foreign financing of national health systems, not whether a truck cleared Kerem Shalom, Rafah, or a Syrian crossing. A higher proxy value is not “more access.” Readers should treat the time series as lineage, not as an operational definition.

Humanitarian access is also distinct from food-price pressure, immunization investment, and forced-displacement scale. Those can move with access or against it. A siege can coexist with high global food prices; an open corridor can coexist with a measles outbreak. Neighbouring drivers are not substitutes.

Mechanism

Access fails or holds through named channels.

First, physical control of routes: parties who hold crossings, ports, airspace, and checkpoints decide whether relief moves. Denial, delay, inspection regimes, and “dual-use” lists are the instruments. A declaration at the UN does not open a gate.

Second, security of the humanitarian worker and of the remaining civilian infrastructure. Bombardment of warehouses, telecommunications cuts, and the collapse of hospitals convert a paper corridor into a dead one. ICRC and WHO incident records are the usual sources, not diplomatic adjectives.

Third, consent, visas, and telecommunications. Even where a party has formally agreed to relief, staff lists, radio frequencies, and fuel allocations can stop the same relief.

Fourth, diversion and taxation. Aid that is looted, taxed, or redirected does not reach the intended civilian. That is an access failure of a different kind from a closed crossing; both belong on this driver.

Fifth, funding and logistics outside the theatre. A pledged dollar is not a delivered litre of fuel. Pipeline breaks at procurement, shipping, and last-mile contractors.

Distribution is the point of the mechanism. The same “access” number can hide one neighbourhood that is reachable and one that is not. Age, disability, internment, and the presence of armed groups change who can walk to a distribution point. An average truck count conceals that.

Timing is short for a closed gate (hours to days) and long for reconstruction of a health system (years). A diplomatic recognition or a Security Council statement can change political cost; it changes calories only if a subsequent, dated access record shows movement.

Evidence base

The best evidence for this driver is operational: OCHA movement statistics, WHO health-cluster updates, ICRC public statements on impediments, and government crossing logs. Those sources name dates, crossings, and denial rates. They do not, by themselves, prove that a distant foreign minister caused the denial.

The live Factrail series currently uses WHO external health expenditure via the World Bank API (https://api.worldbank.org/v2/country/all/indicator/SH.XPD.EHEX.CH.ZS?date=2020:2025&format=json&per_page=20000 (opens in a new tab)). Across the stored 2020–2023 observations the normalized values run from 0.00240682 (2020) to 0.00307636 (2021) to 0.00291067 (2023). Those figures are real draws from that series. They are not a measure of siege or of corridor throughput. Context for genuine access reporting sits at OCHA (https://www.unocha.org/ (opens in a new tab)) and the WHO Health Cluster, not in the health-financing ratio.

Evidence is strongest for a named crossing on a named day. It is weaker for the claim that a diplomatic status change improved access. It is weakest where the only attached number is a global health-finance share. Correlation between that share and under-five mortality is not causation and is not access.

Effect on welfare indicators

Humanitarian access, when it actually moves, is a first-order channel into survival indicators.

• Under-five mortality: direction lower when access holds, higher when it collapses, through nutrition, infection, and lost obstetric care. Horizon: weeks to a few years. Confidence: medium where the access failure is documented in the same population as the deaths; low when the only link is a global series. A French recognition speech is not this channel.
• Measles vaccination coverage: direction higher when cold chain and campaigns can move, lower when they cannot. Horizon: months to years. Confidence: medium for a documented campaign halt; low as a global inference.
• Internal displacement: access collapse often accompanies, and sometimes causes, further displacement; displacement also destroys access. Direction is therefore two-way. Horizon: days to years. Confidence: medium for battlefield-adjacent records, not for a diplomatic communiqué.

New indicator links should be added only with a source that names the channel, lag, and population. Missing coefficients stay missing.

Interactions and contested points

Access interacts with armed-conflict intensity, food-price pressure, immunization systems, and sanctions. Reinforcement is plausible when bombardment and a closed crossing hit the same neighbourhood. Offsetting is plausible when a local ceasefire opens a road while national funding collapses.

The strongest competing view of the current Factrail series is that a health-finance proxy is too remote to carry this driver at all, and that the node should wait for an OCHA-based series. That objection is serious.

Established: the legal and operational meaning of humanitarian access; the distance between that meaning and SH.XPD.EHEX.CH.ZS; the 2020–2023 proxy observations already stored. Not established: a universal coefficient from diplomacy to trucks, or from this proxy to child mortality. A later OCHA series, or a dated crossing log tied to a Fact, would change the graph. Until then the driver is a mechanism with a weak meter, not a score.