A road is a simple thing. In Nepal, it is also hard to pin down on paper. The national budget may show one line for road construction, but underneath it sit 2,110 separate projects. This way of grouping things hides a blunt truth: governments at every level are good at setting aside money for infrastructure but very poor at actually using it. The gap between what is promised and what gets built is now holding back the country’s development.
For decades, public investment has been anaemic. Capital spending averaged a mere 2.7% of GDP in earlier periods, well below regional peers. A surge followed the devastating earthquake of 2015, peaking at 11.4% of GDP in the 2021 financial year. That proved to be an aberration. Between 2020 and 2024 total capital allocations across all levels of government fell from 19.8% to 11.9% of GDP.
More telling than the size of the budget, however, is the failure to use it. The federal government’s rate of capital budget execution—the proportion of allocated funds actually spent—fell from an average of 74% between 2012 and 2019 to roughly 60% in the past five years. This chronic under-execution has cost the country an average of five percentage points of GDP in lost public investment a year since 2021. The consequence is a public capital stock that has dwindled from about 75% of GDP in the mid-1990s to just 54% in 2019.
The failure begins at the very conception of a project. Rules for appraising and selecting projects exist on paper but are weakly enforced. The National Project Bank was established to serve as a central clearing house, vetting proposals before they enter the budget. Its guidelines, revised as recently as March, require pre-feasibility studies and cost thresholds.
Yet it lacks standardised templates and clear methodologies, rendering it ineffective. Projects frequently find their way into budgets without proper scrutiny. This wishful planning is compounded by the Medium-Term Expenditure Framework, which tends to overstate the fiscal space available in future years. The fallout is a portfolio of commitments that is fundamentally unaffordable. Consider 17 ongoing “national pride” projects. At current allocation rates, they would take 41 years to complete. Dedicating the entire federal capital budget for 2024 to them would still require eight years.
Even when a project is funded, it is rarely ready to build. Critical upstream bottlenecks such as land acquisition and obtaining tree-cutting clearances cause long delays. Many projects reach the implementation stage with incomplete designs or vague cost estimates, a prelude to certain overruns. The procurement process is slow and skewed by an overreliance on selecting the lowest bid, which is certain to lead to poorly qualified contractors and further delays. Once under way, projects are hampered by rigid budget reallocation rules and cash-management problems, which stifle the flow of funds. High turnover among project staff and weak performance management undermine continuity and accountability.
This is systemic dysfunction. Fiscal decentralisation, which devolved powers to provincial and local governments, fragmented capacity without ensuring a smooth transfer of skills or data. Information systems do not talk to one another; the project bank and the ministry budget system operate in separate silos, forcing manual and error-prone consolidation of data. Political incentives remain perverse. Grand announcements of new projects provide visible signs of action, whereas the less glamorous work of completion and maintenance is neglected.
The scale of waste is big. Execution shortfalls represent billions of dollars in lost investment that could have fostered growth, created jobs and improved public services. The overcommitted portfolio crowds out essential maintenance, raising the long-term cost of infrastructure and shortening its useful life. It also saps confidence among private investors and citizens alike.
Fixing this means a sequence of practical reforms. The government must enforce its own rules, refusing to fund any project that has not been properly vetted by the National Project Bank. Data systems must be integrated with unique project identifiers to allow for real-time, transparent tracking. The project portfolio needs rationalisation, with a freeze on new commitments until the existing backlog is under control. Readiness bottlenecks, particularly for land, must be simplified. Procurement should move away from a narrow focus on the lowest bid to include quality assessments.
Finally, building institutional capacity and retaining skilled project teams are essential for lasting change. Without such measures, Nepal’s infrastructure budgets will remain little more than elaborate fictions, and its roads will stay on the drawing board. ■