Paradox #52
A Maintenance Budget Labelled Transformation
→ The national budget is not yet a transformation machine. It is mostly a national maintenance machine — paying debt, schools, hospitals, civil-service salaries, and basic infrastructure. The transformation agenda runs on different vehicles.
Bhutan's FY 2025–26 debt servicing (interest + principal)
Nu 28.33 billion
Budget Report; Nu 10.91B interest + Nu 17.43B principal
Bhutan's FY 2025–26 capital budget for "transformation" initiatives (AI, GMC infrastructure, private-sector catalysis, hydropower acceleration combined)
a small fraction of debt servicing
The full numbers
Bhutan’s official narrative talks about transformation: AI, digital economy, Gelephu Mindfulness City, energy abundance, 21st Century Economic Roadmap, USD 5 billion GDP by 2029, USD 100 billion GDP by 2050, GNH-aligned development, regional connectivity, tourism scaling, manufacturing. Bhutan’s actual FY 2025–26 budget (per the MoF Budget Report) tells a different story:
| Budget category | FY 2025–26 (Nu B) | Share of total |
|---|---|---|
| Debt servicing (interest + principal) | ~28.3 | ~32% |
| Current expenditure (salaries, services, administration) | ~45–50 | ~50% |
| Capital expenditure (infrastructure, equipment) | ~30–35 | ~35% |
| Of which: explicitly “transformation” (digital, GMC support, AI initiatives, etc.) | ~3–5 | ~3–5% |
| The mismatch is structural: a budget weighted toward maintenance and debt cannot finance transformation at the scale the narrative implies. |
Imagine this
A senior MoF officer reviews the FY 2025–26 budget request from a line ministry. The ministry has asked for Nu 800 million to launch a national AI literacy programme across schools. The officer scans the available envelope. After salaries (which must be paid), debt servicing (which must be paid), social services (which must be paid), capital maintenance (which must be paid), and ongoing FYP commitments (largely committed), the discretionary envelope for “new initiatives” is small — perhaps Nu 200–400 million across the entire ministry. He approves Nu 80 million. The programme will launch as a pilot in 12 schools. The remaining 554 schools (paradox #17) will wait. Meanwhile the public narrative says Bhutan is “transforming through AI.” Both statements — the small allocation and the transformation rhetoric — are true simultaneously. The gap between them is the unspoken structural reality of how a small-population state with high debt and universal social services actually budgets.
Where this came from
Bhutan’s social-welfare model (free healthcare, free education, near-universal subsidies) was a deliberate Royal Government commitment from the 1970s. It delivered remarkable outcomes for a low-income country. But it locks ~50% of recurring budget into current expenditure that cannot be cut without breaking the social compact.
Debt servicing has grown as hydropower debt accumulated — now ~70% of total public debt is hydropower-related, INR-denominated, and on rigid amortisation schedules. Servicing cannot be deferred without sovereign-rating consequences. The remainder — capital expenditure — is largely committed to infrastructure already in the FYP pipeline (roads, schools, hospitals, hydro project equity contributions).
The truly discretionary fiscal space for “new initiatives” is a few percent of total budget.
Why this matters now
The decade ahead requires investments that don’t fit in the traditional budget envelope:
- GMC infrastructure: USD 5–10 billion over the next decade
- Hydropower 13th FYP pipeline: USD 26 billion through 2040
- Digital infrastructure for AI / data: USD 1–2 billion
- Climate resilience + adaptation: USD 2–3 billion
- Diversified industrial base + private sector catalysis: USD 1–2 billion Total: USD 35–45 billion over 15 years — against a total annual government budget of ~USD 1.5 billion.
The arithmetic does not work through the traditional budget. The transformation will require new capital vehicles: sovereign investment funds, project finance, FDI, multilateral concessional finance, diaspora capital, public–private structures, capital markets, and sovereign reserve deployment (including FX strategies discussed elsewhere).
What it should be
A two-track fiscal architecture:
- Track 1 (the existing budget): continues funding salaries, services, debt, maintenance, and the social compact — the national maintenance function
- Track 2 (a sovereign transformation vehicle): funded through DHI, SOEs, FDI, sovereign bonds, RMA reserve deployment, diaspora capital, multilateral concessional finance, and project-finance structures — the national transformation function Track 2 should NOT be funded out of the normal budget. The normal budget is full. Track 2 is built outside the annual MoF appropriation cycle, with different governance, different risk-tolerance, and different time horizons (10–30 years vs annual).
How others do it
- Singapore — Temasek + GIC as transformation vehicles separate from the operating budget; sovereign wealth deploys long-term capital
- Norway — Government Pension Fund Global (oil revenues) funds long-term investments; operating budget kept disciplined
- UAE (Abu Dhabi, Dubai) — ADIA, Mubadala, ICD as transformation vehicles; emirate budgets separate
- Saudi Arabia — PIF restructured 2015–2020 as the primary transformation vehicle for Vision 2030; outside MoF budget cycle
- Bhutan: DHI exists but is sized for SOE oversight, not transformation-scale capital deployment; the gap is not yet bridged
The question we should be sitting with
If the national budget is already consumed by debt servicing, salaries, social services, and maintenance — where will the money for transformation come from? And if the answer is “outside the budget,” who is designing that vehicle?