What does French public debt actually cost?

France is a useful case well beyond France itself. It is a large advanced economy where a general mechanism became unusually clear: for three decades the debt stock grew while the cost of carrying it fell, so the burden stayed quiet — and then, within a few years, the scissor began to close. What happens when that reversal arrives is now playing out in public, on a scale large enough to make the mechanism visible.

To the question “what does the debt cost”, the most direct answer is not the size of the stock — it is the interest paid each year by general government: 66.6 billion euros in 2025, or 2.2% of GDP and 4.3% of all public revenue (Eurostat). That burden is up 124% from the exceptional low of 2020 — and 80% from 2019, before the pandemic. Two baselines, one conclusion.

The stock itself stands at 3,536.1 billion euros in Q1 2026, or 117.5% of GDP.

The chain has four links. A state continuously refinances maturing securities, so borrowing heavily does not by itself make the stock grow — what makes it grow is the deficit — that is, net government borrowing — plus stock-flow adjustments. That stock carries an average financing cost. Each year, that cost gives rise to interest payments: an obligation already contracted, rather than a new discretionary spending choice, and one that reduces the room left for other choices. And when market rates climb, that average cost follows with a lag — the annual bill swells, and the question stops being how much and becomes who will pay it.

The scissor: thirty years of anaesthesia, then the turn

The observation. From 1995 to the turn of the 2020s the gap between the two curves opens like a pair of scissors: the stock climbs from 57.8% to over 100% of GDP, while the interest burden falls from 3.5% to 1.3% (2020, 29.7 billion euros). For more than two decades, the falling cost of finance let the burden shrink in proportion to GDP even as the stock kept growing.

The mechanism. The link between the two curves is the effective interest rate on the debt — a year’s interest divided by the stock outstanding at the start of that year: a good proxy for the average cost of the stock — not an exact measure of it, because it compares an accounting flow of interest with a stock measured at a single point in time, while the two differ in scope, timing and the treatment of indexation. Nor should it be confused with the rate at which France borrows today.

This is the curve that receives the least attention in public debate, and the missing piece needed to make sense of the rest. The burden does not depend on it alone, but on the product of the stock and this average cost. For thirty years the fall in one offset the rise in the other — which is why doubling the stock did not double the bill. Today the stock is far heavier and the average cost is climbing: both terms now push in the same direction — from about 6.5% in 1996 to 1.2% at the trough of 2020, and back to 2.0% in 2025. Transmission is delayed: a given year’s interest bill covers debt issued at many different dates, so the average cost can keep rising even after market rates settle, as cheap old debt is refinanced.

The reading. Since 2022 — as inflation, index-linked bonds and monetary normalisation have converged — the scissor closes: rising rates meet a debt stock roughly twice as large relative to GDP, and the burden reaches 66.6 billion euros (2.2% of GDP) in 2025. Within the framework of anthropy, this sequence can be read as a cost displaced in time and then returning: displacement, saturation, return. That reading is set out in the working paper AWP-07 — The anthropic loop and applied to debt in AWP-03.

One point often overlooked in public debate: over the INSEE series available since 1995, the highest ratio of debt to GDP remains that of Q1 2021 (117.8%, at the height of the pandemic). It is in current euros that recent quarters set new records.

What 60.1 billion euros in interest represented in 2024

The clearest measure of the burden on public finances: in 2025, interest paid amounted to 4.3% of all public revenue (1,561.6 billion euros of revenue, Eurostat). For comparison, using 2024 — the most recent year for which comparable Eurostat expenditure-by-function data are available:

• Law courts: 8.3 billion euros — in the sense of the European COFOG classification (item GF0303), not the total budget of the French Justice ministry — the interest burden is about 7.2 times that amount;
• Public order and safety, the whole item (GF03): 52.1 billion euros — interest exceeds the entire category;
• Education (GF09): 148.6 billion euros — interest represents about 40% of it;
• Health (GF07): 261.2 billion euros — interest represents about 23%.

These comparisons are about scale, not causation. They show the size of annual debt-service costs relative to public resources — a weight, not a slice taken from another budget.

What the data does not show

The aggregates are unambiguous: no fall in health or education spending. In 2024, both are stable or rising, in euros and as a share of GDP. Anyone claiming that debt has “already cut” those budgets is saying more than the data does.

Hence an apparent paradox: if budgets rise, why do hospitals, schools and courts seem starved? The most common explanation — plausible, but not demonstrated by the series on this page — rests on two standard mechanisms. Public services rely heavily on human labour: their costs track wages, not the productivity gains of machines (Baumol’s cost disease, well established in the economics of services). Demand may also grow faster than GDP: ageing and costly medical progress in health, litigation in justice. If both mechanisms hold, a spending category that remains stable as a share of GDP does not guarantee a stable volume or quality of service. Establishing that it actually fell would require what this page does not measure: sectoral inflation, wages, productivity, demographics, and the volumes actually delivered. Perceived deterioration and rising aggregate spending are therefore not necessarily contradictory: both can coexist if the gap between needs and resources is widening.

Debt did not create that gap — it squeezes the room that would close it: the 66.6 billion of interest paid in 2025 reduce, every year, the fiscal space available to close that gap, all else equal — which is not the same as saying that every euro of interest is a euro taken from health or education. The question the data poses is therefore not “are budgets falling?” but “who will absorb the adjustment” as debt service climbs: higher taxes, cuts to other spending, wider deficits, inflation, or future generations.

Where these figures come from

Every figure on this page is derived automatically from official sources, never copied by hand: quarterly Maastricht debt from INSEE (series 010777616 — stock in billions of euros — and 010777608 — % of GDP); interest paid by general government and total revenue (Eurostat, gov_10a_main, D41PAY and TR); expenditure by COFOG function (Eurostat, gov_10a_exp); the effective interest rate computed as a year’s interest over the stock at the end of the previous year. The official series are re-queried every month; the date below is updated only when an official release changes a figure — most recent data update that changed a value: 17 August 2026. The consolidated data is published openly as dette_officielle.json under a CC BY 4.0 licence — free reuse, including commercial use, on the single condition that the source is cited. The raw series belong to INSEE and Eurostat; what is licensed here is the compilation: the assembly of series, the derived quantities (effective rate, ratios, single-vintage equivalences) and their reconciliation.

The framework used to interpret these figures is set out in What is anthropy? and in the book ANTHROPY — A Big History of Civilization’s Hidden Costs. The book-length treatment of French public debt, Dette Publique : Qui paie vraiment ?, exists in French only.