Debt
A debt is a sum borrowed that must be paid back, with interest. The debt of the French State and the rest of general government reaches 115.7% of GDP (the wealth produced in one year) at the end of 2025. This explanation covers what a country's debt is for, what makes it dangerous, the role of the borrowing rate, and why a lot of debt, or none, can help or hold back its growth.
Last updated: 26 October 2026
In short: a debt is neither good nor bad in itself; it all depends on what it finances, on the rate paid compared with growth, and on the direction it is taking. The rest of this explanation shows why.
What a debt is
A debt is a sum borrowed that will have to be paid back. The lender, the creditor, generally expects to be paid for lending: each year the borrower pays interest, a percentage of the amount owed, until the loan is repaid. A household buying a home on credit, a firm financing a machine, a government building a railway line: all of them can borrow.
This explanation is mainly about public debt: what general government has borrowed and not yet repaid. In France, general government means the State and national public bodies (the State alone carries about 82% of that debt in mid-2026), social security and local authorities (municipalities, departments, regions).
Two words not to confuse:
The deficit: what is missing over one year, when spending exceeds revenue. It is covered by borrowing.
The debt: the total of these loans, piled up year after year, minus what has been repaid.
The deficit is like the water running from the tap during one year; the debt, like the water collected in the bathtub.
At the end of 2025, France's public debt stands at €3,460.5 billion. To judge how heavy it is, it is compared with gross domestic product (GDP), the wealth the country produces in one year: the taxes that pay the interest are levied on this production. The debt equals 115.7% of GDP, a little under fourteen months of the country's entire production (115.7% of 12 months). This ratio, the debt divided by GDP, is what is called the weight of the debt.
Who lends to France
To borrow, a government mostly sells bonds, which are debt securities: the buyer pays it a sum today; the government pays interest every year, then repays the sum on a set date, the maturity. In France, these are French Treasury bonds (OAT, obligations assimilables du Trésor), issued by Agence France Trésor. Every year, Parliament votes, in the budget act (loi de finances), how much the State plans to borrow.
At the end of 2025, according to an IMF estimate, a little over half of the public debt (52.7%) was held by investors based outside France. The rest was held by French players: the Banque de France, banks, insurers and investment funds. Part of the savings households put into life insurance can thus help finance the public debt.
A government does not repay its debt the way a household repays a loan. When a bond reaches maturity, it repays the sum, most often with a new loan. This is called rolling over the debt: the lender is indeed repaid, but the debt itself is renewed instead of disappearing. Unlike a household, a government has no planned end of life: it can carry a debt without ever bringing it down to zero, as long as its lenders trust it and the interest remains bearable. The useful question is therefore not "can everything be repaid?", but "is the weight of the debt in the country's wealth rising or falling, and at what cost?".
How debt works
Public debt can speed up a country's growth or slow it down. In general, this is how it works.
1. The government borrows. By selling bonds, the government gets a sum today that it does not raise in taxes. It can thus spend more than its revenue for the year, or keep up its spending when its revenue drops, for example during a recession.
2. The money is spent or invested. What the loan finances decides what comes next. Part of it goes to current spending: civil servants' pay, social benefits, running public services. Another part goes to investment: roads, networks, buildings, equipment, research. Current spending supports activity at the time. Well-chosen investment can, in addition, raise for years what the country is able to produce. The line is not always clear-cut: a teacher's salary is current spending, but it also prepares tomorrow's production.
3. Interest is paid every year. In general, every euro borrowed costs interest, every year. In 2025, the interest burden of French general government reached about €65 billion, or 2.2% of GDP. This money pays the lenders: it can go neither to other spending nor to lowering taxes. The higher the debt and the rates, the bigger this share grows.
4. The weight of the debt depends on a race. This weight is the debt divided by GDP: it rises when the debt grows faster than GDP, and it falls in the opposite case. To see the race on its own, imagine that the government's revenue equals all its spending except interest: we say it has no primary deficit. It then borrows each year just enough to pay its interest: its debt rises by the amount of that interest, that is, at the pace of the average interest rate it pays. GDP, for its part, grows at the pace of nominal growth, which combines the rise in production and the rise in prices. Prices matter: inflation swells GDP measured in euros, while for most of what the government borrows, the amount to be repaid is fixed in advance, in euros.
Comparing the speed of the debt with that of GDP therefore comes down to comparing the rate with nominal growth. If growth outpaces the rate, the debt weighs less and less relative to the wealth the country produces, without being repaid. If the rate outpaces growth, it weighs more and more, even without a primary deficit. This is why an investment that raises growth can help twice: it adds output, and this higher GDP lightens the weight of the debt.
Inflation acts on both sides of the race. On the one hand, as we have just seen, it immediately swells GDP in euros: after Covid, the rebound in activity and the surge in prices lowered the weight of France's debt. On the other, if it persists, it pushes central banks to raise their rates, and new borrowing by the government generally costs more. But most past borrowing keeps its rate until it matures: the average rate on the debt therefore rises only gradually, as old loans are replaced and new ones are added. The simple calculation below shows this race with numbers.
A simple calculation
The amounts are chosen to be easy to follow. A country has a debt of €100 billion and a GDP of €100 billion: its debt equals 100% of GDP. It pays 3% interest, or €3 billion, which it borrows; all its other spending is covered by its revenue. One year later, its debt is therefore €103 billion.
If its GDP grows by 2%, inflation included, it reaches €102 billion, and the debt weighs 103 / 102, or about 101.0% of GDP. Its weight has risen.
If its GDP grows by 4%, it reaches €104 billion, and the debt weighs 103 / 104, or about 99.0% of GDP. Its weight has fallen, without any repayment.
Same debt, same effort: only the gap between the rate and growth changed the direction.
When it does not go as planned
The investment does not pay off: spending that is poorly chosen, or that produces nothing more, leaves the debt without the growth that was supposed to lighten it.
Cutting the deficit too fast in the middle of a crisis: lowering spending or raising taxes slows activity. During the crisis that followed 2008, the advanced economies that had planned the largest deficit reductions grew less than forecast, especially early in the crisis (Blanchard and Leigh, 2013). If growth weakens, the weight of the debt may fall less than expected, or even rise.
Repaying at any cost: in the 1980s, Ceaușescu's Romania chose to repay all its foreign debt, what it owed to foreign lenders, as fast as possible; it managed to do so in 1989, at the cost of a recession and of sharply reduced consumption (Ban, 2012).
Borrowing rates in brief
This section gives the essentials; the explanation "Borrowing Rates", to be read next, details them. The borrowing rate is the interest rate a borrower pays: the price of borrowed money. For a government, it is set on the market: investors buy its bonds, and the price they are willing to pay sets the rate: the lower the price, the higher the rate. For example, a bond that pays €3 of interest a year and repays €100 at maturity yields 3% if you pay €100 for it, and more if you pay only €95: you receive the same €3 for a smaller outlay, and get €100 back at maturity. This rate depends mainly on the central bank's rates (the ECB for the euro), on the rates expected for the years ahead, on the length of the loan and on lenders' confidence, that is, the risk of non-repayment they perceive.
On 1 January 2026, France borrowed over 10 years at 3.58%, Germany at 2.88%: the gap, 0.70 point (3.58 − 2.88), is the extra that lenders ask of France compared with Germany. This gap matters: one extra point of interest, once the whole debt has been renewed at that rate (which takes years), would cost about €35 billion of interest a year (1% of €3,460.5 billion). In mid-September 2026, amid a general rise in rates, the gap exceeded 0.90 point, as did Italy's.
The rate on a new bond is not the rate on the whole debt. In 2025, the public debt as a whole cost on average about 2.0% a year, less than market rates at the time, because a large part of it had been borrowed when rates were low. The State's securities reach maturity, on average, in 8.5 years: each year, only part of them is renewed at the current rate, along with the new loans that finance the deficit. A rise in rates is therefore passed on slowly, but it lasts: according to the government's forecasts, as a result of both higher rates and a larger debt, interest would rise from 2.2% of GDP in 2025 to 3.7% in 2030.
The explanation Borrowing Rates takes up these three points in the same order: how the market sets the rate, what makes it rise or fall, including lenders' confidence, of which the gap with Germany gives an idea, and how a rise reaches the budget, year after year.
What debt makes possible
Investing without waiting: a railway line, an electricity network or a university serve for decades. Financing them by borrowing spreads their cost over the generations that benefit, instead of making the taxpayers of a single year pay the full price. This is the logic of the rule imposed on French local authorities: they may borrow to invest, but not to finance their operating expenses nor to repay their loans (article L1612-4 of the general code of local authorities).
Absorbing a shock: in a recession, tax revenue falls and some spending rises, such as unemployment benefits. Borrowing makes it possible to avoid cutting spending at the worst moment. In France, during the Covid crisis, public debt rose from 98.2% of GDP at the end of 2019 to 114.9% at the end of 2020, partly because GDP itself fell.
What makes it dangerous
It costs every year: when the debt and rates rise, interest takes a growing share of the budget.
It can grow by itself: if the rate paid persistently exceeds nominal growth, the weight of the debt rises even without a primary deficit. This is the snowball effect.
It makes the country dependent on confidence: a heavily indebted country must borrow a lot each year to renew its debt. If lenders doubt it will be able to pay, they ask for a higher rate; the debt then costs more, which can deepen their doubts. Nor can a euro area government decide to create money to pay its lenders. It borrows in euros, a currency it does not create, and the European treaties forbid the ECB and the national central banks to lend to governments or to buy their bonds directly from them, that is, when governments sell them (article 123 of the Treaty on the Functioning of the European Union); they can only buy them on the market, from other investors, and under conditions. When confidence vanishes, the way out can be brutal: in 2012, Greece carried out the largest public debt restructuring in history, that is, an agreement to repay less than planned; its private creditors lost more than half the value of their securities (Zettelmeyer, Trebesch and Gulati, 2013).
It reduces the room for the next crisis: the higher the debt, the more a new shock pushes it towards levels that worry lenders.
It can hold back private investment: a government that borrows a lot can push rates up for everyone, and firms then invest less. This is the crowding-out effect; it is generally weak when rates are very low.
A lot of debt or none at all
Neither high debt nor the absence of debt is good or bad in itself. Three questions matter most: what the debt finances, the rate paid compared with growth, and the direction it is taking. European rules use a reference value of 60% of GDP for debt, but this figure is not a proven danger threshold. A 2010 study, Reinhart and Rogoff's, presented further down, had concluded there was a threshold at 90% of GDP; more recent studies, including the IMF's in 2014, found no clear one.
High debt can be useful when it has financed investments that pay off, when growth exceeds the rate paid and when lenders keep their confidence. It can become harmful when it mainly finances current spending outside crises, when the rate exceeds growth, or when it leaves the country with no room to face the next shock. It can then slow growth in two ways: through the interest, which ends up forcing tax rises or spending cuts, and through the higher rates it can bring for everyone.
The absence of debt avoids interest and leaves room for crises. But a country that refuses all borrowing must pay cash for every investment: it risks making fewer of them, or later, and its future growth may suffer. A less indebted country may also choose to borrow to invest: in March 2025, Germany, whose debt stood at 61.4% of GDP at the end of 2024 against 112.6% for France, amended its constitution, the Basic Law, to create a €500 billion fund, financed by borrowing and meant for investment, mainly in infrastructure, over twelve years.
What the studies say
Optional on a first reading.
In 2010, Carmen Reinhart and Kenneth Rogoff gathered more than 3,700 annual observations on 44 countries over about two centuries. They found that growth excluding inflation varies little with debt as long as debt stays below 90% of GDP, but that beyond that, median growth (the middle value when the observations are ranked from lowest to highest) falls by about one point, and average growth considerably more.
In 2013 (article dated 2014), three researchers from the University of Massachusetts, Thomas Herndon, Michael Ash and Robert Pollin, went back over the same data. They found errors in their spreadsheet, excluded data and a questionable way of computing the averages. Once these errors were corrected, the advanced economies whose debt exceeded 90% of GDP had, from 1946 to 2009, average growth of 2.2% a year excluding inflation, not −0.1% as the 2010 study had reported: not very different from that of less indebted countries. According to them, the 2010 result had been widely used, after the 2008 crisis, to support austerity policies.
In 2014, an IMF study covering 34 advanced economies found no clear threshold beyond which growth deteriorates sharply; countries whose debt was high but falling grew as fast as the others (Pescatori, Sandri and Simon).
When high debt and weak growth go together, the link can run both ways: a recession itself pushes debt up, because revenue falls and support spending rises. The fact that the two go together does not, on its own, prove that debt holds back growth.
France in figures
The figures below come from the Court of Audit. They are in points of GDP (one point equals 1% of GDP), and the change in the weight of the debt splits into three pieces: the gap between the average rate paid and nominal growth, the primary deficit, and other operations that change the debt without going through the deficit (for example when the State borrows to lend: its debt rises, without this being spending). The first piece cannot be read directly from the gap between the two rates: the Court of Audit calculates it from this gap and the weight of the debt the year before, with more precise figures than the rounded ones shown here; the calculation is detailed in the explanation "Fiscal Policy".
In 2024: nominal growth (3.6%) clearly exceeded the average rate (1.9%), and this gap took off about 1.7 points; the primary deficit added 3.7 and the other operations 1.1. In total, −1.7 + 3.7 + 1.1 = 3.1 points: the debt went from 109.5% to 112.6% of GDP.
In 2025: nominal growth (1.9%) and the average rate (about 2.0%) were almost equal, and the gap hardly played any role (+0.1 point); the primary deficit added 2.9 points and the other operations 0.1. In total, 0.1 + 2.9 + 0.1 = 3.1 points: the debt went from 112.6% to 115.7% of GDP.
Despite the fall in the primary deficit and in the other operations, the rise was the same as in 2024, because nominal growth (1.9% instead of 3.6%) no longer offset the interest.
To remember
Most important:
Debt and deficit: the deficit is what is missing over one year; the debt is the accumulated total of loans not yet repaid.
The right yardstick: debt is measured against GDP, because it is the country's production that makes it possible to pay the interest.
What it finances: borrowing to invest can raise future growth; borrowing for current spending that does not prepare the future, outside crises, mostly leaves the bill.
The race between rate and growth: if nominal growth exceeds the rate paid, this gap lowers the weight of the debt; in the opposite case, it pushes it up, even without a primary deficit. A primary deficit pushes the weight of the debt up further; a primary surplus lowers it.
Confidence: an indebted government depends on its lenders; when they doubt, the rate rises and the debt costs more.
Neither too much nor none: high debt reduces room for manoeuvre and exposes the country to crises, but the absence of debt can deprive it of useful investments, and repaying it too fast can make it poorer.
Optional on a first reading:
The average rate and today's rate: a rise in market rates weighs on interest only little by little, at the pace at which old bonds are replaced and the deficit is financed.
No magic threshold: studies more recent than Reinhart and Rogoff's (2010), including the IMF's in 2014, find no level of debt beyond which growth deteriorates sharply; the direction the debt is taking also matters.
Sources
ReferencesThis explanation lays the foundations. The calculation of the change in the debt from one year to the next, and its breakdown for 2025, are detailed in the explanation "Fiscal Policy" (ch. 6, Intermediate level); the debates on the burden of debt and its political causes, in "The Political Economy of Public Debt" (ch. 15, Advanced level).
Works and sources cited in this explanation:
INSEE (May 2026): "Le compte des administrations publiques en 2025", Insee Première no. 2106 (in French).
French Treasury and Agence France Trésor (October 2026): Rapport sur la dette des administrations publiques, annexed to the 2027 draft budget bill (in French).
Court of Audit (June 2026): La situation et les perspectives des finances publiques (in French).
Eurostat (October 2026): general government gross debt as a percentage of GDP, series gov_10dd_edpt1.
Reinhart and Rogoff (2010): "Growth in a Time of Debt", NBER Working Paper no. 15639, revised December 2011.
Herndon, Ash and Pollin (2014): "Does High Public Debt Consistently Stifle Economic Growth? A Critique of Reinhart and Rogoff", Cambridge Journal of Economics, 38(2), pp. 257-279.
Pescatori, Sandri and Simon (2014): "No Magic Threshold", Finance & Development, IMF, June 2014, summary of their working paper "Debt and Growth: Is There a Magic Threshold?" (IMF Working Paper no. 14/34).
Blanchard and Leigh (2013): "Growth Forecast Errors and Fiscal Multipliers", NBER Working Paper no. 18779.
Zettelmeyer, Trebesch and Gulati (2013): "The Greek Debt Restructuring: An Autopsy", Economic Policy, 28(75), pp. 513-563.
Ban (2012): "Sovereign Debt, Austerity, and Regime Change: The Case of Nicolae Ceausescu's Romania", East European Politics and Societies, 26(4), pp. 743-776.
Bundestag (March 2025): report of the sitting of 18 March 2025 on the reform of the debt brake (in German).
French ministry in charge of local authorities: Principes du recours à l'emprunt (in French).
Market impact
MarketsThe rate at which France borrows is set on the government bond market, by supply and demand. When investors doubt the path of the debt, they demand a higher yield, and the spread over Germany, the euro area benchmark, widens. For France, this spread was 0.70 point on 1 January 2026 and exceeded 0.90 point in mid-September 2026, as did Italy's, amid a general rise in rates (debt report annexed to the 2027 draft budget bill). Rating agencies sum up their assessment of the risk of non-repayment in a rating. A downgrade can force insurers, pension funds and investment funds to sell some securities, because holding rules rely on these ratings.