How the IMF Measures Government Debt — and Why Figures Disagree

Look up "Italy's government debt" in three different places and you may get three different numbers. None of them is necessarily wrong. Government debt statistics are not a single, universally agreed figure — they're the output of a set of methodological choices, and different institutions make different choices for defensible reasons. Understanding those choices is the difference between reading a debt ranking as a fact and reading it as what it actually is: a number produced under a specific, statable definition.

Gross debt versus net debt

Gross debt counts everything a government owes, full stop — every bond and loan outstanding, regardless of what assets the government also holds. Net debt subtracts the government's financial assets — cash, foreign-exchange reserves, holdings in sovereign wealth or pension funds, loans the government itself has made — from that gross figure. For a country with few financial assets, the two numbers are close. For a country that has accumulated a large asset pool, they can diverge enormously. Norway's government has borrowed to run efficient domestic capital markets even though it is, on a net basis, one of the world's largest creditor governments once its oil fund is counted; Japan's gross debt is roughly 230% of GDP, but a meaningful share of that is offset by financial assets the state itself holds. Gross debt is the more commonly cited figure internationally, mainly because it's more consistently reported and easier to compare across countries — not because it's inherently the more meaningful one.

General government versus central government

A second major choice is which layer of government gets counted. Central government debt covers only the national government itself. General government debt — the basis the IMF's headline comparisons and this site generally use — adds in state, provincial, and local governments, along with social-security funds, giving a fuller picture of the public sector's total obligations. In a federal system with heavily indebted states or provinces, or in a country whose local governments borrow significantly, the general-government figure can run well above the central-government one. China is the clearest illustration of how large this gap can get: its narrowly defined general-government debt looks moderate by international standards, but once local-government financing vehicles — the off-books borrowing entities used to fund infrastructure — are folded in under an "augmented" measure the IMF has published, the effective figure is meaningfully higher. That gap is large enough that it functions as its own case study in how measurement choices shape headline conclusions (see our explainer on China's hidden LGFV debt).

Market value versus face value

A third, more technical difference concerns how outstanding bonds themselves are valued. Face value (also called nominal or par value) counts each bond at the amount the government promised to repay at maturity. Market value instead uses what those bonds would currently sell for, which moves with interest rates: when rates rise, the market value of older, lower-yielding bonds already outstanding falls below face value, and vice versa. National debt-management offices sometimes report market-value figures for their own operational purposes, but most cross-country rankings — including the IMF's and this site's — use face value, since it's simpler, more stable, and doesn't require constantly re-marking the entire stock of debt.

Currency and consolidation choices

Two further wrinkles compound the picture. Debt denominated in foreign currency has to be converted into a common reporting currency, usually US dollars, for cross-country comparison — and exchange-rate swings alone can move a country's reported debt figure from one year to the next even if nothing about its actual borrowing changed. And "consolidation" — whether debt one government entity owes to another part of the same government is netted out or counted twice — varies by methodology. The IMF generally consolidates general-government debt so that, say, a country's national government borrowing from its own social-security fund isn't double-counted as if it were two separate liabilities; not every national statistical agency applies the same consolidation rules to its own domestic reporting, which is one more reason a country's own treasury figure and its IMF-reported figure can diverge even when both are being produced honestly and competently.

Why Eurostat, the IMF, and national treasuries disagree

Layer these definitional choices on top of differing accounting standards, differing treatment of state-owned enterprises and public guarantees, differing fiscal-year timing, and differing revision schedules, and it becomes clear why the same country can show up with different debt-to-GDP figures in the IMF's World Economic Outlook, in Eurostat's Maastricht-criteria debt figures for EU countries, and in a national treasury's own domestic reporting. None of these is "the" correct number in some absolute sense; each is correct under its own definition, and the definitions exist for different institutional purposes — Eurostat's Maastricht figures exist specifically to police EU fiscal rules, for instance, which shapes what it does and doesn't include.

Timing and revisions also matter

A final, less-discussed source of disagreement is simply timing. Government debt figures are typically reported with a lag of months, sometimes revised repeatedly as more complete data comes in, and published on different schedules by different institutions — the IMF twice a year with its World Economic Outlook, national treasuries often monthly or quarterly, Eurostat on its own cycle tied to EU fiscal-surveillance deadlines. A snapshot taken from one source in one month can look meaningfully different from a snapshot taken from another source a few months later, even before any methodological difference comes into play, simply because the underlying GDP denominator itself gets revised as national accounts data matures.

What this site uses

This site's live counters and rankings are built on general government gross debt, sourced from the IMF's World Economic Outlook database, generally the April 2026 vintage, cross-checked against World Bank and national-treasury figures where available. That's a deliberate methodological choice, not the only defensible one: it favors comparability across a large number of countries over precision on any single country's net fiscal position. Readers who want a country-specific deep dive on assets versus liabilities — Singapore, Norway, and Japan being the clearest cases — can find that context in our companion piece on when government debt isn't what it seems. Full detail on sourcing, vintage, and update cadence is documented on the methodology page.