Two official Belgian government numbers can differ without either one being wrong. The difference may come from what is being measured, who is included, when the transaction is recorded, which budget stage or source vintage is used, or whether the statistic has since been revised.
The useful question is therefore not simply “Which number is correct?” It is: What exactly does each number represent?
The short answer
Official figures are claims made inside a measurement framework. If two sources use different frameworks, stages or vintages, they can legitimately produce different numbers for something that sounds similar in ordinary language.
For Belgian public-finance data, six distinctions recur often enough that Cameralius treats them as part of the fact itself:
- institutional perimeter — which public bodies are included;
- fiscal stage — budget, commitment, liquidation, payment or another execution state;
- budget or source vintage — initial, adjusted, provisional, revised or final;
- accounting basis and timing — for example budget-accounting treatment versus ESA accrual recording;
- reference period or frequency — annual, quarterly, calendar-year or another reporting period;
- statistical revision state — whether later source information or methodological improvements have changed the published estimate.
Different number ≠ contradiction.
A contradiction exists only when two claims are supposed to describe the same thing under the same relevant conditions and cannot both be true.
1. The institutional perimeter can be different
A federal budget figure is not automatically comparable with a Belgian general-government figure. Belgium’s political structure contains the Federal State, Communities, Regions and local authorities, while ESA general government uses the statistical sector S.13 and its four subsectors.
The National Bank of Belgium publishes government-finance statistics for S.13 as a whole and for federal government, Communities and Regions, local government and social-security funds separately. A figure covering only the federal subsector can therefore differ substantially from a figure covering all S.13 even when both are labelled “Belgian government expenditure”.
See what counts as Belgian government for the institutional-perimeter rules.
2. The fiscal stage can be different
A budget authorization is not the same observation as a commitment, liquidation or payment. Belgium’s federal accounting framework distinguishes those stages rather than treating them as one generic form of “spending”.
That means a €10 million budget amount and €8 million of later budgetary expenditure can both be correct. They answer different questions: what was authorized, and what reached the relevant execution stage.
See why a Belgian federal budget is not the same thing as spending and why budgetary expenditure does not necessarily mean the money was already paid.
3. The budget vintage can be different
One fiscal year can have an initial budget followed by one or more approved adjustments. BOSA’s federal archive shows multiple adjustments in individual years, including two expenditure-budget adjustments in 2024 and three in 2022.
An initial figure and a later adjusted figure can therefore both be official and correct while describing different authorized states of the same fiscal year.
See why Belgium’s initial and adjusted federal budgets can show different numbers.
4. The accounting basis or timing can be different
Official budget accounts and national-accounts statistics are built for different purposes. Belgian government-finance statistics produced under ESA 2010 are harmonised statistical accounts, not simply copies of administrative budget tables.
The National Bank explains that source data from government are verified, corrected and harmonised to meet ESA 2010 standards. Eurostat’s methodology adds an important timing rule: ESA records transactions on an accrual basis—when the economic activity takes place or the claim or obligation arises—rather than simply when cash is paid.
So an administrative cash or budget-accounting figure and an ESA expenditure figure can differ even if they relate to the same underlying policy activity.
5. The reference period can be different
Annual and quarterly statistics do not necessarily have the same level of detail, publication timing or revision state. Eurostat publishes government-finance statistics annually and, at a lower level of detail, quarterly. The National Bank likewise produces Belgian national accounts on different publication schedules.
A quarterly value, an annual total and a rolling-year percentage can all be valid but should not be compared as if they were the same observation. The reference period and frequency belong in the label.
6. The statistic may have been revised
Official statistics are not always final when first published. Eurostat explicitly describes revision as a normal part of producing high-quality European statistics: new source information, corrections or methodological improvements can change previously published estimates.
The National Bank says Belgian national-accounts estimates are likely to be revised over time as source information becomes available and that, as a general rule, figures only become final after roughly four years. A first official estimate and a later revised estimate can therefore both have been correct official publications at the time they were released.
This does not mean truth is relative. It means the evidence state changed.
A worked example
Imagine you find three official figures for “Belgian government spending on X”:
- €100 million in the initial federal budget;
- €115 million in a later adjusted federal budget;
- €108 million in an ESA general-government statistic published after the year ended.
Those numbers do not automatically conflict. The first two describe different budget vintages. The third may use a broader institutional perimeter and an accrual-based national-accounts methodology rather than the federal administrative budget framework.
Before deciding that one number is wrong, you would need to align the perimeter, fiscal stage, vintage, accounting basis and reference period.
What this does not mean
Official sources can contain mistakes. Two figures can genuinely contradict each other. A revision can correct an error, and a source can be incomplete or badly labelled.
The point is narrower: difference alone is not evidence of error. You must first test whether the two figures are actually comparable claims.
A practical comparison checklist
When two official Belgian figures differ, Cameralius asks:
- Are they about the same institution or statistical perimeter?
- Are they at the same fiscal stage?
- Are they from the same budget or data vintage?
- Do they use the same accounting and timing rules?
- Do they cover the same reference period and frequency?
- Has either series been revised since the other was published?
- Are the units, prices, classifications and consolidation rules the same?
If those conditions are not aligned, the numerical difference may be meaningful rather than erroneous.
The Cameralius rule
Cameralius will not resolve conflicting-looking figures by silently choosing the newest, largest or most convenient number. The system should first identify the semantic dimensions that make the observations comparable—or explain why they are not comparable.
Where revisions occur, earlier vintages should remain historically legible. Where definitions differ, those differences should remain visible. A clean answer should come after the evidence has been aligned, not before.
Related Cameralius research
- A Belgian federal budget is not the same thing as spending
- Why Belgium’s initial and adjusted federal budgets can show different numbers
- Does Belgian federal budgetary expenditure mean the money was actually paid?
- What counts as Belgian government?
- Zero is not missing in Belgian public data
The broader evidence rules are collected in the Cameralius methodology. You can also explore the Belgium research hub.
Sources & evidence
- National Bank of Belgium — Public finances methodology. Defines S.13 and its subsectors and explains how Belgian source data are harmonised to ESA 2010.
- National Bank of Belgium — National accounts: general. Explains the Belgian publication sequence and that estimates are revised as more complete source information becomes available.
- Eurostat — Government finance statistics methodology. Explains ESA 2010 classification and accrual timing for government transactions.
- Eurostat — Data revision policy. States that revision is a normal process in European statistics and distinguishes routine, major and unscheduled revisions.
- Eurostat — Government finance statistics: information on data. Documents annual and quarterly availability, the S.13 perimeter and revision handling.
- BOSA — Le cycle budgétaire. Describes Belgium’s federal budget stages and adjustment process.
- BOSA — Budget initial and Budget ajusté. Preserve different authorized budget vintages for the same fiscal year.
Evidence review: 4 September 2026. This article synthesizes several distinct Belgian and European measurement frameworks; the metadata of the specific source remains controlling in any concrete comparison.
Editorial status
Country: Belgium
Publication type: Explainer
Topic: Evidence, comparability & revisions
Last evidence review: 5 September 2026