In June 2026, the average rate on new housing loans in France was 3.16%. That same month, on exactly the same loans, it was also 3.73%.
Both figures are accurate, both were published by the same institution, and they do not measure the same thing. The first is an interest rate. The second is a price. The whole difficulty, for a borrower, is that only the first is on display everywhere.
The gap between the two barely moves: 0.61 of a percentage point on average over the first seven months of 2026 — that is, roughly 20% more in annualised cost. This is not a freak month, it is a constant.
What this article covers
Why the rate the whole market talks about is not the one you pay, and what that gap comes to in euros on an average loan. Which rate actually protects you, and which one binds nobody. And why the 35% rule is not a law — which changes what you can negotiate.
Two rates, one population
The figure picked up by every rate tracker comes from the same place: the Banque de France’s monthly statistics on lending to households. What those trackers leave out is the label the Banque de France itself puts on its own chart: “Rate (excluding fees and insurance) on new housing loans”.
The methodological note is blunter still: “In accordance with the harmonised ECB definitions, the rates recorded are effective rates in the narrow sense (TESE); they correspond to the interest component of the Taux Effectif Global.” In other words, this figure measures what the bank charges by way of interest, and nothing else.
The second figure exists, and it is public
Under the European framework for interest rate statistics, the Banque de France reports two twin series to the European Central Bank for housing loans to households: the annualised agreed rate — the interest component — and the annual percentage rate of charge. Same definition of the population, same month, same collection. Here are the two series for 2026.
| Month | Nominal rate | TAEG | Gap |
|---|---|---|---|
| January 2026 | 3.05% | 3.65% | +0.60 pt |
| February 2026 | 3.10% | 3.71% | +0.61 pt |
| March 2026 | 3.10% | 3.71% | +0.61 pt |
| April 2026 | 3.11% | 3.73% | +0.62 pt |
| May 2026 | 3.10% | 3.71% | +0.61 pt |
| June 2026 | 3.16% | 3.73% | +0.57 pt |
| July 2026 | 3.18% | 3.80% | +0.62 pt |
How to read this: the July figure for the TAEG is provisional as we write. The six preceding months are final. The gap moves between 0.57 and 0.62 of a point: it is stable, which rules out any statistical artefact.
⚠️ Do not compare these figures with mortgage brokers’ rate trackers. The European series includes renegotiations; the most widely quoted French statistic excludes them, which gives 3.27% in June instead of 3.16%. Setting a nominal rate that excludes renegotiations alongside a TAEG that includes them would inflate the gap artificially. The two columns above come from the same series, differing by a single character in its identifier.
What the gap costs, in euros
The average loan granted in France in the first four months of 2026 was €191,000, excluding renegotiations, with an average initial term of 22 years and 8 months. Let us apply June’s two rates to it.
| A €191,000 loan over 272 months | Monthly payment | Total cost of the credit |
|---|---|---|
| At the nominal rate (3.16%) | €984 | €76,741 |
| At the actual TAEG (3.73%) | €1,041 | €92,266 |
| Difference | €57 / month | €15,525 |
That €15,500 is not a hidden surcharge: it is the fees and the insurance, expressed as a rate equivalent. Making them comparable is precisely what the TAEG is for.
What the TAEG contains, and what it leaves out
Article R. 314-4 of the code de la consommation (the French Consumer Code) lists what goes into the TAEG: arrangement fees, fees paid to intermediaries, the cost of compulsory insurance and guarantees, account-keeping and payment-instrument charges, and the cost of valuing the property — in so far as they are necessary in order to obtain the credit, or to obtain it on the terms advertised. Article R. 314-5, by contrast, excludes the costs of acquiring the property, transfer duties and notary’s fees included.
Borrower’s insurance therefore counts, but only where it is a condition of the loan being granted — which in practice it almost always is. It is by far the largest building block of the gap, and it is also the most negotiable: we have set out what the law allows in our article on switching your mortgage insurance.
What the gap is made of
Three items make it up, in that order of importance. Insurance first: 92.6% of the outstanding stock of housing loans is covered against death, and 76% against incapacity for work. Then the guarantee: two thirds of the outstanding stock is backed by a guarantee institution and a quarter by a mortgage charge; a mortgage charge carries a land registration tax of 0.715% of the sums secured and a property security contribution of 0.05%, whereas a guarantee is paid for in another way. Arrangement fees last, weighing the least.
To that must be added, where applicable, the intermediary’s remuneration. And this is where a structural asymmetry comes into play: point 2° of article R. 314-4 places within the TAEG “fees paid or owed to intermediaries involved in any way whatsoever in the granting of the loan” (our translation). A mortgage broker charges the client 1 to 1.5% on average, and additionally receives 0.25 to 0.5% from the bank. Communicating on the TAEG rather than on the nominal rate would mean, for a broker, displaying their own remuneration in the headline figure. It is easy to see why the market compares itself on the other one.
One consolation, all the same: set against its neighbours, French credit remains cheap. In September 2025, the average rate on new mortgage loans stood at 2.99% in France, against 3.32% on average in the euro area, 3.28% in Italy and 3.74% in Germany. Five euro-area countries did better: Malta (1.89%), Spain (2.66%), Finland (2.78%), Slovenia (2.83%) and Portugal (2.86%). To which is added a French feature that amounts to protection: 99.6% of 2025 lending is at a fixed rate, whereas Spanish or Italian borrowers often bear the interest rate risk themselves.
The only rate that protects you
There is a statutory ceiling on the cost of credit, and it does not bear on the advertised rate. Article L. 314-6 of the code de la consommation defines a usurious loan as one granted “at an overall effective rate which, at the time it is granted, exceeds by more than one third the average effective rate applied during the preceding quarter”.
So it is the TAEG, fees and compulsory insurance included, that is compared with the threshold. An application can show a perfectly reasonable nominal rate and run into the usury ceiling once insurance is added — the classic situation for an older borrower, or one presenting a health risk.
Usury thresholds applicable from 1 July to 30 September 2026, for fixed-rate mortgage loans: 4.07% for a term of under 10 years, 4.57% from 10 to under 20 years, 5.29% from 20 years upwards. Variable-rate loans are at 5.28% and bridging loans at 6.39%. ⚠️ These thresholds are quarterly: the fourth-quarter figures appear in the Journal officiel at the end of September and apply from 1 October. Always check the current quarter before signing.
The 35% rule is not a law
You read everywhere that “the law imposes a maximum debt ratio of 35%”. The claim is wrong on three counts, and each one changes what a borrower can hope for.
What it really is, and who it is addressed to
It is a decision of the HCSF (Haut Conseil de stabilité financière — the French macroprudential authority), taken on the basis of point 5° of article L. 631-2-1 of the code monétaire et financier (the French Monetary and Financial Code), which empowers it to “set conditions for the granting of credit” (our translation). It is therefore not a recommendation — indeed the same article opens up an action for annulment before the Conseil d’État (France’s supreme administrative court), the hallmark of a regulatory act.
But that decision is addressed to credit institutions, supervised by the ACPR (Autorité de contrôle prudentiel et de résolution — the French banking and insurance supervisor). Not to borrowers. A loan granted beyond the threshold remains perfectly valid: no provision makes it void. And the HCSF itself writes that its decision “can in no way take the place of” the banks’ risk policy, the use of the departure allowance being a matter for their “free assessment” (our translation).
The allowance exists, and it is not saturated
The decision authorises banks to depart from the criteria on 20% of their quarterly lending. In the first quarter of 2026 they used 17.5% of it — a record since the measure was created, but still below the cap. The average debt-service ratio (taux d’effort) actually observed is 30.6%, for an average maturity of 22.6 years: the average borrower is not at the cap, but more than four points below it.
Two points that circulate in garbled form: the allowance bears on 20% of quarterly lending, not on “20% of applications”; and the usury rate has not been revised monthly since 1 January 2024, contrary to what a great many pages still claim.
The debt-service ratio, as the HCSF defines it
The official definition is broader than the one usually read, on three counts. In the numerator it takes in all loans, including those taken out with other institutions, principal, interest and compulsory borrower’s insurance included. In the denominator it takes the household’s net income before tax. And it is assessed not over the first year but over the highest annual level recorded across the whole term of the loan — an application at 36% for five years only does not comply.
That last subtlety explains many a refusal the borrower never understood. To place your own capacity, we set out the orders of magnitude in our article on the salary needed to borrow €200,000 to €300,000.
A confusion to clear up: the law that was never withdrawn
Since the spring, a great many pages have claimed that a private member’s bill was due to abolish the 35% rule and that it was “withdrawn during the night of 29 to 30 April 2026”. These are two different texts, two years apart, systematically conflated.
The text actually withdrawn is a private member’s bill from the sixteenth parliamentary term, tabled in January 2024 and withdrawn by its author on 30 April 2024. The 2026 text, tabled on 14 April, has never moved: its file contains nothing but a referral to the finance committee. One clue ought to have raised the alarm — 29 April 2024 was a Monday, as the page of the Assemblée nationale, the lower house of the French parliament, states; 29 April 2026 is a Wednesday.
The framework itself has not changed: the HCSF maintained it in June 2026. A parliamentary report from the same period recommends raising the allowance from 20 to 25% and building in a notion of residual living income; those are recommendations, with no normative force as things stand.
Mon Simulateur Immobilier monthly payment calculator
The calculator gives you the monthly payment and the total cost of your credit from the amount, the term and the rate, and lets you compare an offer at the advertised nominal rate with the same offer at the TAEG shown on the loan offer. That gap, and that gap alone, measures what you actually pay.
Conclusion
The credit market compares itself on a rate that is not a price, because it is the one that gets negotiated and the one that makes the headlines. There is nothing scandalous in that: the statistic is correctly labelled, and serious players do mention the caveat. But the caveat is at the foot of the page, and the figure is in the headline.
The practical consequence is simple. When you compare two offers, do not compare nominal rates: compare the TAEGs, which must appear on every loan offer, and the total cost in euros. Run both scenarios on your own case with the Mon Simulateur Immobilier monthly payment calculator. On the average loan, the gap between the two readings is worth over fifteen thousand euros.
