Nepal has 77 districts. The largest INGOs operating in the country are, in any given year, present on the ground in roughly thirty of them. The remaining forty-seven receive either no INGO presence at all or are touched only by short-term project work that arrives, delivers, and leaves on the same fiscal year. This is not an accident of geography. It is a structural feature of how the development sector organises itself in Nepal, and once you see the pattern it is hard to unsee.
The data
Drawing on the Association of International NGOs in Nepal’s published member registry, cross-referenced against IATI-published project location data and the Social Welfare Council’s project approvals database, the picture of district-level INGO presence in mid-2026 looks like this.
Eleven districts — Kathmandu, Lalitpur, Bhaktapur, Kaski, Chitwan, Morang, Sunsari, Banke, Rupandehi, Kavre, and Makwanpur — host the headquarters or a permanent field office of more than ten Tier-1 INGOs each. These are the country’s urban centres and their immediate hinterlands. They contain roughly one-third of Nepal’s population and absorb, by our best estimate from project disclosures, between fifty-five and sixty per cent of total INGO programme spending.
Twenty more districts — a band of mid-hill and Terai districts that includes most of the well-served humanitarian and post-earthquake reconstruction targets — host between three and ten Tier-1 INGO offices each. These contain roughly another third of the population and absorb roughly thirty per cent of programme spending.
Forty-six districts host fewer than three Tier-1 INGO offices on a permanent basis. Within this group, twenty-two districts host none. These contain the remaining third of the population and absorb the remaining ten to fifteen per cent of programme spending.
The arithmetic is uncomfortable. The bottom third of districts by INGO presence receive, per capita, roughly one-fifth of the resources that the top third receive.
Where the gap is
The under-served districts are not randomly distributed. They cluster in three places.
The first is the upper Karnali — Mugu, Humla, Dolpa, Jumla, and Kalikot. These districts rank consistently in the top decile of Nepal’s national vulnerability indices: highest child stunting, lowest female literacy, lowest road density, highest infant mortality. They also rank consistently at the bottom of INGO presence rankings. The disconnect is the starkest single mismatch in the data.
The second is the eastern hills — Sankhuwasabha, Solukhumbu (away from the Everest tourism corridor), Khotang, Bhojpur, and Okhaldhunga. These districts are less acutely vulnerable than upper Karnali but are similarly absent from the active INGO footprint, and have been since the 1990s.
The third is the central Terai — particularly the cluster of Bara, Parsa, Rautahat, Sarlahi, and Mahottari. These are some of Nepal’s most populous districts, with significant Madhesi and Muslim minority populations, persistent poverty pockets, and a long-running pattern of being treated as transit corridors rather than implementation sites. They are heavily under-served relative to population.
Why the pattern persists
The structural explanation is in three parts.
Operational logistics. Running an INGO field office in upper Karnali costs roughly three to four times what the same office costs in the Kathmandu Valley or a Terai district. The drivers are predictable: per diem rates for staff travel, helicopter contingency budgets, security and risk management overheads, the cost of running diesel generators when grid power fails. When donors enforce strict cost-per-beneficiary ratios, the upper Karnali offices lose those calculations almost regardless of programmatic logic.
Donor preferences. The major bilateral and multilateral donors are not, in fact, indifferent on geography. FCDO has explicit Karnali and Sudurpaschim priorities. The World Bank’s federalism support work concentrates in provincial capitals. The EU funds urban governance disproportionately. The vector sum of these preferences is a heavy weighting toward provincial centres and a structural under-weighting of the eastern hills and the central Terai.
Implementing-partner inertia. Most INGOs entered specific districts in specific years for specific programme reasons, and have stayed in them long past the original logic. The map of INGO presence in 2026 is, in significant part, a fossil of the map of INGO presence in 2004 and 2015 — the major Maoist conflict and post-earthquake reconstruction inflection points. New districts are entered slowly; old districts are exited even more slowly.
A specific example
Consider the contrast between Kavre and Sankhuwasabha. Both are mid-hill districts. Kavre, an hour east of Kathmandu, hosts the field operations of fourteen Tier-1 INGOs. Sankhuwasabha, in the eastern hills bordering Tibet, hosts two. Kavre’s population is roughly 380,000; Sankhuwasabha’s is roughly 160,000. On a per-capita basis Kavre receives perhaps fifteen times more INGO programme attention than Sankhuwasabha — and there is no obvious vulnerability-based reason for that ratio. Kavre’s child stunting rate is meaningfully lower; its road density is dramatically higher; its proximity to Kathmandu means that any technical support, supervision visit, or evaluation costs a fraction of what the same activity costs in Sankhuwasabha.
The reason Kavre is over-served is, principally, that it is easy. A staff member based in Kathmandu can run a site visit and be back the same evening. Sankhuwasabha requires a Twin Otter or a two-day road trip. The system rewards what is easy to administer over what most needs administering.
What the federalism transition was supposed to fix
The 2015 constitution’s federal structure was designed, among other things, to push resources and decision-making closer to underserved populations. In a programmatic sense the constitution has worked. The local government grant formula, set by the National Natural Resources and Fiscal Commission, applies an explicit weighting for human development indicators and geographic remoteness. Money flowing through the state system is now more redistributive than it has ever been in Nepal’s history.
Money flowing through the INGO system is not. The local government grants and the development sector’s project spending are governed by entirely different logics, and they pull in opposite directions on geographic equity. The constitutional reform that pushed state resources outward has, perversely, made the inward-pulling weight of donor-funded programming more visible by contrast.
What this means
The coverage gap is not a finding that will surprise anyone working in Nepal’s development sector. It is one of the field’s open secrets. What is worth noting is how reliably the gap reproduces itself year to year. Without an active intervention — and there is no current donor coordination initiative in Kathmandu explicitly targeting the geographic redistribution of INGO presence — the same map will be true in 2027 and in 2030.
The argument for engaging with this is not principally one of equity, although equity is the most defensible framing. It is operational. The districts where INGO presence is thinnest are also the districts where local-government capacity is weakest and federal-transfer absorption is lowest. They are exactly the places where the kind of patient, multi-year institutional accompaniment that INGOs are best at providing would deliver the most marginal value. The system is, in this sense, badly allocated against its own theory of change. The data has been saying this for years. The donor-coordination architecture has not yet found a way to respond.