The 99.9%
Budget

Liberia spent US$879,596,767 against an FY2025 appropriation of US$880,661,873 — an execution rate of 99.9%. Read as a headline, that is a government that spent almost exactly what the legislature told it to.

Read by sector, it is not. Execution ranged from 68.6% to 108.0%, and US$21.7 million — 2.5% of the national budget — moved between sectors during the year. The sectors that lost it were the ones the FY2025 budget had promised the largest increases.

Sections 1 to 5 are drawn from the Ministry of Finance and Development Planning's Annual Consolidated Budget Performance Report FY2025, on the on-budget appropriation basis; section 6 adds the Comptroller and Accountant General's accounts for the same year. Sector figures are Appendix 3's throughout; the report disagrees with itself by one dollar on two of them, which is logged as D10 and noted below. Sources and limits are set out at the end.

01The aggregate rate is true, and tells you nothing

Four sectors spent US$21.68 million more than they were appropriated; seven spent US$22.75 million less. The two nearly cancel, which is why the total looks clean.

Public Administration alone accounts for 68% of all the overspending, and with Security and Rule of Law, 92% of it.

Two-thirds of this movement was later written into the government's own final budget; section 6 sets the two budgets side by side.

02The sectors promised the most delivered the least

The FY2025 budget was a 19.2% increase on the FY2024 recast, and it was not spread evenly. Setting what each sector was promised against what it received:

Across the eleven sectors, the correlation between the size of the promised increase and the execution rate is −0.90. Dropping the two extreme cases — Agriculture, which had by far the largest increase, and Social Development Services, the only sector cut — it is still −0.63.

The result does not depend on which FY2024 figure the increase is measured from, with one exception. The FY2024 column here is the recast budget the report itself compares against. Measured from FY2024's outturn, printed since in the FY2026 budget book, the correlation is −0.91, and −0.70 without the two extremes. Only FY2024's budget as first enacted weakens it: −0.73, and nothing once the extremes go. That is the allocation the government later recast (D8).

Eleven observations cannot establish a mechanism, and the direction is not surprising: a sector whose budget more than doubles has more new activity to stand up, and new activity is harder to execute than a payroll. But the effect is large enough to matter for how the budget should be read.

The larger the announced increase, the less of it arrives — while the large administrative and security votes not only hold their increases but add to them in-year.

It is the size of the increase that matters, not the size of the sector: the two smallest sectors, Industry and Commerce and Social Development Services, both spent about 108% of their budgets.

Agriculture is the clearest case. Its appropriation was raised 147.9%, from US$6.75m to US$16.75m — the government's most emphatic budget-line statement of priority in FY2025. It spent US$11.49m. Its investment programme, US$8.00m appropriated, disbursed 49.9%, and none of it as capital: the report's own sector table puts Agriculture's spending on non-financial assets at US$7,553 against US$8.08m appropriated (Table 20, p. 41).

03Almost half of Public Administration is debt service

The largest sector in the Liberian budget is also the least well named. All of the government's debt-service payments — domestic and foreign liabilities alike — are booked inside it.

Debt service is 44.9% of the sector and 17.4% of the entire national budget — larger than the whole health appropriation (US$91.3m) and larger than the whole education appropriation (US$119.7m). Netting it out does not rescue the execution rate: the non-debt half of Public Administration overran by US$12.50 million, or 6.7%.

Within the debt line, the performance report has the two halves moving in opposite directions: domestic liabilities paid at 112.4% (+US$9.68m), particularly to the Central Bank of Liberia, and foreign at 90.2% (−US$7.39m) after a debt restructuring (pp. 31–33). That US$17 million swing is the report's own split, and the government's other counts do not share it. The Comptroller and Accountant General's accounts divide the same US$155.29m into US$80.08m domestic and US$75.22m foreign — 102.8% and 100.1% of their appropriations (p. 17) — and the report's own debt chapter puts domestic debt service at US$80.18m (Table 36, p. 66). The two documents agree on the total, not on the split.

04The swing line is goods and services

Table 10 (p. 33). Debt service is in section 3; subsidies and recurrent capital, US$3.5m and US$7.1m, are not shown.

Use of goods and services overran by US$11.30 million. Outside debt service, where domestic payments ran at 112.4%, it is the only large recurrent class that did, and it is concentrated:

Public Administration's recurrent goods and services alone overran by more than the government-wide total. Health and Education went the other way on their recurrent budgets: they spent US$4.04 million less on goods and services than appropriated — 86.0% in Health and 89.9% in Education, while their payrolls ran at 99.5% and 98.8%.

That was not a cut to what the two sectors bought. Both spent most of their investment budgets on goods and services as well, and the report's own sector tables, which count both budgets, put goods and services at 116.5% of appropriation in Health and 120.0% in Education — on drugs and medical supplies, the report says, and on teaching and learning materials (Tables 16 and 18, pp. 38–40). What both sectors left unspent was capital: non-financial assets executed at 10.7% in Health and 7.4% in Education. Behind sector rates of roughly 98%, the investment budgets were spent on goods and services instead.

The report attributes the Public Administration overrun to "vendor share" — payments to service providers that deliver public services under revenue-sharing agreements, which rose with revenue — the ECOWAS trade levy, debt, Liberia's UN Security Council representation, an assets-recovery task force, and additional support to the Liberia Revenue Authority in response to "financing pressures arising from the drawdown of USAID support" (p. 34).

05The capital programme, again

The Public Sector Investment Plan executed at 93.6% — US$99.85m against US$106.71m. Differencing the report's total and recurrent economic classifications shows what that money bought.

Appropriated as 100% capital, executed as 62.5% capital. Measured against the capital appropriation, real capital delivery was 58.5%, not 93.6%.

The sector data sharpens this. The two most capital-dependent sectors are Infrastructure and Basic Services (81.0% of its budget is PSIP) and Agriculture (47.8%) — and they are the two worst executors in the budget, at 88.1% and 68.6%. Across the eleven sectors the correlation between PSIP dependence and execution is −0.63. The final budget went further, and cut capital spending as a whole by more than a third (section 6).

06Against the final budget, the spread falls from 39 points to 8

Everything above measures FY2025 against the budget as enacted, which is how the performance report measures it. The Comptroller and Accountant General's annual accounts for the same year print a second budget beside it — the final budget — by sector (Table 56, p. 63) and by economic class (p. 17). Its total is the same, US$880.66 million. The allocation inside it is not.

The final budget raised the four sectors that overspent by US$14.3 million — two-thirds of their US$21.7 million overspend — and took the same amount, net, from the seven that underspent. Agriculture was cut by 30.4%, from US$16.75m to US$11.65m; Infrastructure and Basic Services by US$4.86m; Health by US$2.17m. Public Administration was raised by US$7.86m and Security and Rule of Law by US$4.58m. In ten of the eleven sectors, the final budget moved the way the spending went; the exception, Transparency and Accountability, was raised by US$0.14m and underspent both budgets.

Against the final budget, every sector executed between 94.3% and 102.0%. The spread that section 1 measured at 39.4 points is 7.7 against the budget as revised.

Enacted budget and outturn: performance report, Table 2 and Appendix 3. Final budget: accounts, Table 56 (p. 63), printed in US$ thousands. The accounts book a further US$20.0m in Public Administration — domestic borrowing passed to the Liberian Bank for Development and Investment as a capital injection — which the performance report does not carry; with it, the sector spent 107.7% of its final budget.

The economic classification shows the same movement in the capital budget. The final budget cut capital spending — the investment plan's and the recurrent budget's together — from US$113.8m to US$74.3m, by more than a third, and raised goods and services from US$149.6m to US$199.1m. The US$70.6m of capital actually spent is 62.0% of the enacted figure and 95.0% of the final one.

Accounts, statement of comparison of budget and actual amounts (p. 17), printed in US$ thousands. Its original-budget column equals the performance report's appropriation for every class.

The accounts give reasons for each sector's adjustment and for every agency moved by 10 percent or more, though not who approved them or when. Several cuts are put down to "delay in request and underutilization" — money moved because it was not being spent — and several increases to named purposes, from extra sittings of the legislature to by-elections in Nimba County (pp. 68–75).

None of it reaches the headline rate. The total never moved, so FY2025 executes at 99.9% against the enacted budget and at 99.9% against the final one. What the final budget changes is the spread: written into it, the reallocation all but disappears. That is why the standard test of budget credibility, PEFA's indicator of expenditure composition, compares outturn with the budget as originally approved — and why sections 1 to 5 do the same.

07Where this leaves the FY2025 story

• Revenue and grants came to US$872.60m against US$879.60m spent, a deficit of US$7.00m. A US$12.5m IMF loan, which the performance report counts in its US$885.10m of "total revenue", left the year US$5.50m in hand.
• Total spending held at the appropriation while roughly US$22m moved between sectors after appropriation, in a consistent direction — from capital-heavy and newly-expanded sectors towards administration and security operating costs.
• Two-thirds of that movement, and a cut of more than a third to the capital budget, were written into the government's final budget, whose total did not change.
• The published 99.9% execution rate is arithmetically correct and analytically empty. The 39-point spread beneath it, measured against the budget as enacted, is the number that describes FY2025.

Method

Figures are extracted from the MFDP Annual Consolidated Budget Performance Report FY2025, on the on-budget appropriation basis, in nominal US dollars as printed.

Sector outturns come from Appendix 3 (pp. 74–75), the only table in the report that splits each sector's spending into recurrent economic lines and its investment programme. The same sector totals are printed in the report's per-sector tables, and FY2025 disbursement by sector is also published in the ministry's Annual Fiscal Outturn Report FY2025 and in the Comptroller and Accountant General's accounts, which section 6 uses. Before use, every sector figure was reconciled to the report's own printed totals and cross-checked against the report's per-sector tables (section 5.5, pp. 35–43); the one-dollar disagreements this surfaced are logged as D10 and D11.

The final budget in section 6 is from the Comptroller and Accountant General's Annual Consolidated Financial Report FY2025, which prints US$ thousands. Its sector figures were reconciled to the table's own printed totals and checked against the report's sector-by-sector notes (pp. 68–75), which restate every final budget; its original-budget column matches the performance report's appropriation for every sector to the thousand dollars it prints.

Every figure in this note traces to a document id and page number. Derived figures — execution rates, variances, correlations, the debt-service netting — are calculations from those printed figures and are labelled as such.

What the source does not support

• Off-budget spending is not here. These are appropriation-basis figures. On the IMF's central-government basis, staff estimate FY2025 expenditure at about US$1,170m (expense US$891m plus capital US$279m) against US$880m on-budget. Capital is where the gap is widest: roughly US$279m of capital spending in the IMF's estimate, against US$70.6m of capital on budget. Most capital investment in Liberia is donor-financed and never enters the sector table above, so "Infrastructure underspent" is a statement about the appropriated programme, not about all infrastructure spending in the country.
• The performance report prints no revised appropriation; the accounts do. The report benchmarks the outturn against the original approved budget throughout, and refers once to "revised appropriations" (p. 34) without printing them. The Comptroller and Accountant General's accounts print a final budget by sector and by economic class, which section 6 uses. They give reasons for the adjustments but not who approved them or when, so section 6 describes what changed, not how.
• The investment programme's split by sector is less reliable than its total. Appendix 3's economic lines are recurrent-only and give each sector's PSIP as one figure. The report's per-sector tables (Tables 12–22) split total spending by class, so each sector's PSIP can be recovered by difference, and each reconciles to Table 11 within US$2. Summed, though, they put US$70.57m of the PSIP into capital, against US$62.45m from Tables 9 and 10. Section 5 uses Tables 9 and 10, whose capital total the Comptroller and Accountant General's accounts and the ministry's outturn report both print; the sector-table figures in sections 2 and 4 are quoted as printed.
• Section 2 measures growth from a recast budget. The FY2024 column is the recast budget the report itself compares against, not an outturn, so its growth rates set a budget against an outturn. FY2024's outturn has since been printed in the FY2026 budget book (p. 33); measured from it, the correlation in section 2 is −0.91 rather than −0.90.
• Two $1 discrepancies exist inside the source. Table 2 (p. 17) and Appendix 3 disagree by one dollar on Public Administration and on Energy and Environment, in offsetting directions. Five sector totals differ from their own component lines by $1–$2. Both are logged (D10, D11); neither is material and neither propagates to any figure above.

Sources

FY2024 outturn by sector (section 2): National Budget FY2026, table 1.9 (p. 33). Comparators: IMF Country Report No. 26/98, Table 4a (p. 42), central-government basis. Every figure above is shown with its source on this page; the underlying tables are available on request.

Revisions

10 October 2026. This edition follows a review of every statement against the source documents. The headline figures stand: 99.9%, the 68.6–108.0% range, the US$21.7 million, the −0.90 correlation, the investment programme's 62.5% capital and the whole of section 6. It corrects:

• The year's balance. Section 7 called FY2025's result a surplus of US$5.50m. The US$885.10m the performance report calls total revenue includes a US$12.5m IMF loan; revenue and grants were US$872.60m, US$7.00m short of spending.
• Health and Education. Section 4 said both sectors protected payroll by cutting the goods and services that buy drugs and textbooks. That holds for their recurrent budgets only. Counting their investment budgets, the report puts goods and services at 116.5% and 120.0% of appropriation; what went unspent was capital.
• Goods and services. Section 4 called it the only large recurrent class to overrun; domestic debt payments, in the same table, ran at 112.4%.
• Debt. Section 3 presented the US$17 million domestic–foreign swing as a fact invisible in the aggregates. The report prints it in its economic tables, and the Comptroller and Accountant General's accounts and the report's own debt chapter split the same payments differently. Its heading said Public Administration is mostly not administration; debt service is 44.9% of it.
• Sources. The method called Appendix 3 the only published source of FY2025 disbursement by sector, and the limits said the investment programme could not be split by sector. The report's sector tables do both, the ministry's outturn report and the accounts also print sector disbursement, and the sector tables disagree with Table 9 about how much of the investment programme was capital.
• Smaller points. Four of the gaps the print edition's section 2 tabulates were 0.1 points off (Energy and Environment −5.8, Education −2.4, Transparency and Accountability −0.8, Industry and Commerce +9.0); Agriculture's budget rose 2.5 times, not three; the pull-quote now says what the correlation measures; Agriculture's investment programme disbursed 49.9% but bought no capital; the report's "vendor share" is revenue-sharing payments, not "vendor obligations"; the IMF figures are staff estimates; the final budget's cut to the seven underspending sectors is net; and the accounts explain adjustments of 10 percent or more, not every one.

3 October 2026. Corrected the correlation in section 2 to −0.90; the first edition, of 28 August 2026, printed −0.91. Replaced the first edition's statement that no revised FY2025 budget had been published: the Comptroller and Accountant General's accounts print one, and section 6, which was new, compares the year with it.