France’s public debt: why pay the markets 4.9% instead of its own citizens?

On October 1, 2026, the yield on France’s ten-year government bond hit 4.92%. You have to go back to 2002 to find it higher. And here is a detail any bond trader would have laughed off ten years ago: France now borrows at a higher rate than Italy.

That same day, a French saver who kept money in a Livret A, the country’s tax-free, government-regulated savings account, was earning 1.7%. One who left it in a checking account was earning nothing, or close to it: household demand deposits pay an average of 0.04%.

Those two numbers should not be able to coexist in the same country. On one side, a state paying top dollar to find lenders. On the other, citizens sitting on hundreds of billions that earn nothing at all. The idea of placing part of the national debt with ordinary savers is already in circulation. I want to turn it into a concrete, costed proposal.

A gap nobody looks at

A 4.9% yield is not some act of God. It reflects a risk premium that keeps climbing as France’s fiscal trajectory becomes harder to read, along with a global competition for savings that has grown fiercer. Back in August, I described how the AI giants now line up at the same window as national treasuries, with trillions in commitments to finance. When two borrowers compete for the same pool of savings, the price goes up.

The move has been brutal: 3.73% in July, 4.23% in September, 4.92% at the start of October. And the state cannot simply wait for better days. Agence France Trésor, the agency that manages the national debt, has to raise a record €310 billion in medium- and long-term debt this year. Every additional point of interest ends up, sooner or later, in debt service, which is to say in taxes.

Meanwhile, French savings sit idle. At the end of May 2026, according to Banque de France monetary statistics, individuals held €486 billion in checking accounts, more than in the Livret A (€548 billion if you count the broader household sector). And that money is not spread evenly. According to the appendices of the 2025 annual report on regulated savings, which also break down demand deposits by balance, the 11% of accounts holding more than €10,000 account for 80% of the total. These are safety cushions that have long outgrown their purpose, waiting only for a reason to move.

Who does France owe money to?

At the end of 2025, 56% of the French state’s tradable debt was held by non-residents, and the share is still rising. Pension funds, asset managers, central banks, sovereign wealth funds: more than half of what France owes is owed abroad, and we don’t even have a precise breakdown of who holds the paper.

There is an asymmetry here worth naming. A foreign fund collecting coupons on French government bonds, known as OATs, pays no French income tax on them. This is not some hidden favor: nearly every sovereign issuer waives withholding tax on foreign creditors in order to widen its pool of buyers. But the effect is real. A French citizen who wants to lend to his own country will see his interest cut by the flat tax on investment income, which rose to 31.4% this year. At the same gross yield, the domestic saver is the worst-treated creditor of his own country’s debt.

Nor does that savings stay put. It goes elsewhere. A good share of what the French invest in the stock market ends up financing the American economy, sometimes without even owning it. The result: our savings feed foreign markets, and foreign markets lend to us at 4.9%. We pay interest on the money we let walk out the door.

The proposal: a national loan indexed to the ten-year rate

The principle fits in a few lines.

The state issues a bond reserved for individual French residents. Term: ten years. Return: the ten-year OAT yield on the day of subscription, minus one percentage point, locked in for the life of the bond and paid annually. Interest is fully exempt from income tax and social charges, just like a Livret A. At the October 1 rate, that works out to 3.9% net.

The bond is issued in periodic tranches of limited size, say twice a year. Each tranche takes the market rate of the moment: there is no figure set once and for all that would turn too generous or too stingy as markets move.

This is not a savings account, though. The money is locked up for ten years. Early withdrawal remains possible, but with a penalty steep enough to discourage it. The only exceptions, redeemed at par with no penalty, are death, disability, and the purchase of a primary residence. These are the same life events already recognized by French home-savings plans and employee savings schemes. Nothing exotic.

The cap is set at €50,000 per person, on a registered bond that cannot be topped back up after a withdrawal. That figure is not arbitrary. It matches the target: checking accounts that hold far more than any emergency fund requires. Above that level, you are dealing with people who already have a broker and know how to buy an OAT on the market. They don’t need the simplicity of this product, and the state has no reason to hand them a tax shelter.

Subscribing would be done through one’s bank or at a public counter, with no broker, no fees, and no jargon.

Finally, the product has a single purpose, written into its design: refinancing existing debt at a lower cost. Making debt cheaper must not become an excuse to take on more of it. I said as much about the 2025 budget: mortgaging the future is not a policy. Interest savings that fund new spending are not savings.

What the state gains

This is where precision matters, because this is where an informed reader would be waiting for me.

Indexing makes the math simple and stable: the state pays one point less than it would cost to borrow the same sum for ten years on the market, whatever the level of rates. Each €10 billion tranche saves €100 million in interest a year, for ten years. To give a sense of scale, €50 billion raised in total would yield, at best, €500 million a year.

At best, because that figure is a ceiling. It is reached only if the money raised was sitting idle in checking accounts and replaces ten-year debt that would otherwise have been sold to non-residents. With a resident who would have bought an OAT anyway, the math flips. At 4.92% and a 31.4% flat tax, that creditor costs the state only about 3.4% once the tax is collected. Paying him 3.9% tax-free means losing half a point instead of saving a full one.

The one-point discount is not pulled out of thin air. It roughly matches the tax wedge, the gap of about one and a half points between an OAT’s gross yield and what remains after the flat tax. The scheme splits that wedge between saver and state: the saver beats the OAT’s net return (3.9% versus 3.4%), and the state beats the gross rate it pays abroad (3.9% versus 4.9%). With no discount, the entire gain would go to the saver; at one and a half points, it would go entirely to the state, and the product would no longer beat a market OAT. One point leaves both sides better off.

One condition must remain explicit: the savings exist only if the bond replaces ten-year issuance. The ten-year is currently the most expensive maturity; since January, Agence France Trésor has borrowed at an average of 3.55% across all maturities. If the money raised were used to avoid short-term issuance, which is already cheaper, the gain would melt away. The point of the scheme is to replace the most expensive maturity, not the average, and that is how Agence France Trésor should manage it.

Against a foreign creditor, on the other hand, the gain is complete. The objection that comes to mind first, “the state is giving up the tax on this interest,” does not hold up: it already collects nothing on coupons paid to foreign funds. Replacing an untaxed foreign creditor with an untaxed French saver costs not a cent in tax revenue. It simply costs less in interest.

The biggest gain, moreover, shows up in no spreadsheet. A saver locked in for ten years does not dump his bonds the day a rating agency downgrades France, or the day a government falls. A base of patient domestic creditors is a shock absorber. It is the difference between a country that is at the mercy of market moods and one that can ride them out.

What the saver gains

For the average French saver, the comparison that matters is not the one made on trading floors. It is the one on his bank statement. Next to a checking account that pays nothing and a Livret A at 1.7%, a 3.9% net return guaranteed by the state pays more than double. At the €50,000 cap, that is €1,950 a year, tax-free, with no risk of losing capital at maturity.

Even the insiders’ comparison works in its favor. An OAT bought on the market today yields 4.9% gross, or about 3.4% after the flat tax. The national loan pays 3.9%, net. The market OAT keeps one advantage, liquidity, but it remains out of reach for most savers. Almost no individual buys OATs directly. The market is intimidating, few banks offer access, and those that bother often prefer to sell in-house products that are far more lucrative for them, like those structured products in which the client is really insuring his bank against market crashes. The national loan removes that barrier: a product anyone can understand, a rate known at subscription, a single guarantor.

It is also worth spelling out what the saver gives up. He forgoes liquidity for ten years. Once his tranche is subscribed, his rate is fixed: if inflation takes off or market rates keep climbing, he will not benefit from the rise, except by subscribing to a later tranche within the cap. That is the price of the guarantee, and everyone should be able to weigh it with full knowledge. The scheme is meant for money not needed in the short term, not for an emergency fund.

The two objections that remain

The first is cannibalization. The money will not come only from checking accounts. It will also come out of the Livret A, which funds social housing, out of the guaranteed-capital funds in life insurance policies, out of term deposits. And the banks will fight back, because demand deposits are their cheapest source of funding. The more attractive the product, the greater the risk, and it carries a fiscal cost: money that leaves a taxable investment stops generating tax revenue. The €50,000 cap limits the effect without preventing it. The scheme only works if its target remains idle money, and that is the standard by which it should be judged.

The second concerns the wholesale market. A large pool of retail debt, illiquid by design, could fragment the OAT market and push up the rate the state still pays on the bulk of its €310 billion in annual issuance. Italy sidesteps this risk by paying its savers the market rate. My bond stays one point below the market in gross terms. It only works if the amounts remain modest relative to the overall financing program, which is one more reason for a cap.

Finally, to avoid any confusion: this has nothing to do with the forced, interest-free loan that some senators tried to impose on wealthy taxpayers during the budget debate. Here, nobody is obliged to do anything. The state offers, the state pays, and everyone chooses.

Italy has done it

This is no utopia. Italy created BTP Valore, a family of government bonds reserved for individual savers, with no fees and a loyalty bonus for those who hold them to maturity. Between June 2023 and March 2026, the program raised just over €112 billion. The most recent issue, in March, brought in €16.2 billion from roughly 522,000 subscribers.

The Italian model is not identical to mine. Rome applies a reduced 12.5% tax rate to its government bonds and pays its savers at market rates, plus a bonus. It pays for loyalty. My national loan makes a different bet: one point below the market in gross terms, offset by the tax exemption, which makes it more rewarding on a net basis, and by simplicity, for an audience that has no access to the bond market anyway. Both approaches share the same conviction: a country’s debt is in better hands with its own citizens than with funds that will sell it at the first sign of trouble.

Taking back our debt

For years, French debt has drifted abroad, line by line, auction by auction. We are told that markets pass judgment, that agencies hand out grades, and that we must bow to both. That is not wrong.

The national loan will not reverse this trend. A few tens of billions, however well placed, will barely move the share held by non-residents. But it is a modest and concrete path: French citizens lending to France, at a fair rate, with no middleman. The interest stays in the country instead of flowing to anonymous portfolios, and citizens are no longer locked out of their own debt.