Before looking at financing, it is worth restating the spending-side shift this report builds on, since it is the reason financing needed to grow so much in the first place. Between 2000-2004 and 2020-2024, development spending's share of the budget rose from 27.2% to 41.8%, while "other goods, services and transfers" — the routine, non-wage, non-interest recurrent spending that funds day-to-day service delivery — fell from 45.7% to 20.1%.

Source: TICGL/TERI analysis of Ministry of Finance, Government Budgetary Operations data (IMF format).

This shift, on its own, is not the problem — most economies moving up the income ladder go through an investment-heavy budget phase. The real question is what paid for the extra 14.6 percentage points of development spending, and on what terms. That is where the next three sections go.

This is the most important chart in this report. In the early 2000s, development spending was largely financed by grants — money from development partners that Tanzania did not have to repay. Today, grants finance only a small fraction of development spending, and the rest is borrowed.

Source: TICGL/TERI analysis of Ministry of Finance financing data (IMF format). 2000-2004 exceeds 100% because grants that period included general budget support beyond project financing alone.

In 2000-2004, grants averaged 108.6% of development expenditure — development partners were, in effect, funding the entire development budget and more, through general budget support. That figure fell steadily and sharply: 74.2% (2005-2009), 42.3% (2010-2014), 12.8% (2015-2019), and just 5.8% (2020-2024). This is not a small drift — it is a near-complete reversal of how Tanzania's development spending is funded, from "mostly free" to "almost entirely borrowed," inside two decades.

Some of this shift reflects a genuinely positive trend: Tanzania's transition toward middle-income status has naturally reduced its eligibility for the most concessional grant aid, and greater self-financing capacity is, in principle, a sign of a more mature economy less dependent on donors. But the pace and scale of the shift mean the cost structure of "development" has changed fundamentally — every shilling of development spending today carries a repayment obligation that the equivalent shilling in 2003 did not.

Debt service — interest payments plus foreign debt amortization — as a share of government revenue traces a striking U-shape over the 26-year period: high in the early 2000s, falling sharply after major debt relief, and climbing back up over the past decade.

Source: TICGL/TERI analysis of Ministry of Finance financing data (IMF format). Non-concessional share is of new foreign development project loans disbursed that period.

Debt service stood at 25.0% of revenue in 2000, fell to as low as 6.6% in 2009 — largely thanks to the Heavily Indebted Poor Countries (HIPC) and Multilateral Debt Relief Initiative (MDRI) programmes that wrote off a large share of Tanzania's earlier debt — and has since climbed back to 22.9% in 2025, briefly touching 26.3% in 2023. In absolute terms, Tanzania is back to devoting roughly the same share of its own revenue to debt service as it did at the start of this dataset.

The crucial difference is the composition of that debt. The debt relieved in the mid-2000s was overwhelmingly concessional — low-interest, long-maturity loans from official development partners. The debt now driving the renewed climb is, on the evidence in Table 3, roughly 40-55% non-concessional in every period since 2011 — commercial or near-commercial terms, with shorter maturities and higher interest rates. Put simply: Tanzania has climbed back to a similar debt-service burden as 2000, but this time without the cushion of concessional terms that made the earlier burden more manageable and, ultimately, forgivable.

Alongside foreign borrowing, government has increasingly turned to domestic financial markets — mainly Treasury bonds bought by pension funds, banks, insurance companies and individual investors — to help finance the budget. This matters for ordinary Tanzanians in a more direct, if less visible, way than foreign debt: money that domestic savers lend to government is money not available to lend to private businesses.

Source: TICGL/TERI analysis of Ministry of Finance financing data (IMF format). "Non-bank borrowing" excludes direct bank-sector financing and captures bond-market issuance to the broader public and institutional investors.

Domestic non-bank borrowing averaged just TZS 16.6 billion a year in 2000-2004. By 2020-2024 it averaged TZS 2,056.8 billion a year — a roughly 124-fold increase, far outpacing the roughly 24-fold growth in total government expenditure over the same span. In other words, government's reliance on domestic debt markets has grown more than five times faster than the budget itself.

This is consistent with — though this dataset alone cannot conclusively prove — a "crowding-out" effect: as government issues more Treasury bonds to fund its own spending, it can push up domestic interest rates and absorb savings that might otherwise flow to private-sector loans, potentially making credit scarcer or more expensive for Tanzanian businesses and entrepreneurs. This would be a genuinely important channel through which the budget's financing structure could hold back private-sector-led development, even while government's own investment spending rises. Confirming the size of this effect would require a dedicated analysis of domestic lending rates and private-credit growth alongside government bond issuance — a natural next step for TICGL/TERI's research agenda.

Everything documented so far describes the terms on which Tanzania now finances its development spending: less free money, more debt, a meaningful share of it non-concessional, and a fast-growing claim on domestic savings. None of it, on its own, tells you whether this is a bad deal. A shilling of non-concessional debt spent on a well-executed, high-return project (a port upgrade that cuts logistics costs economy-wide, a power plant that ends load-shedding for manufacturers) can still be a good investment, even at a higher interest rate. The same shilling spent on a delayed, over-budget, or under-utilised project is a bad deal regardless of how cheap the financing was.

This is the genuine limitation of budget-composition and financing-structure data: it can tell you, with precision, that the risk profile of Tanzania's development financing has increased. It cannot tell you, on its own, whether the returns have increased to match. That requires project-level information this dataset does not contain — completion rates, cost overruns, usage/capacity-utilisation data for completed infrastructure, and the growth or revenue actually attributable to specific projects.