Comment Mélenchon compte annuler la dette française

0:00So , just go ahead and grab them and set

0:01them on fire . There are no more .

0:03It was the end of June 2026. Jean-Luc

0:05Mélenchon proposed burning the head of

0:07the French state , which owes its own

0:09central bank , the Banque de France , 488

0:12billion euros . Since then , the country

0:14has been talking about nothing else .

0:16The governor of the Banque de France

0:17came out of his shell to respond on the

0:19radio in three words : “ It’s

0:20illegal , it’s dangerous and it’s

0:22useless . ”

0:23And for the past 6 weeks , economists

0:25and columnists have been talking about

0:26this topic . Meanwhile , 43 % of the French

0:30say yes , cancel it , against 31 % who say

0:32no . It cuts across all parties , and the

0:35election is in 6 months . So today ,

0:37we’re going to take both camps

0:39seriously , one after the other , with

0:40their figures . Those who want to burn

0:42have solid arguments and you’re going

0:44to hear them in full . Those who refuse

0:46also say no , and you won't watch this

0:48from afar . From now on , you are the

0:50finance minister , you are the one who

0:51signs , and at the end , you will know

0:53exactly what you are signing , including

0:55on a point that neither side is putting

0:57forward . Absurd , illusory or even

1:03voodoo , these are the initial reactions

1:05to the proposal . But adjectives are not

1:08arguments . It obviously only convinces

1:10those who already agreed . The proposal

1:13convinces . A Hugov poll from August 31

1:16to September 3 finds 43 % of French

1:18people favor canceling part of the debt

1:20, while 31 % are opposed . The idea goes

1:23far beyond a single political camp . So

1:25why is it so popular ? Because the

1:27French have been told about the

1:28disaster for 20 years and it has not

1:30happened . On September 21 , 2007 ,

1:32François Fillon said that he was at

1:34the head of a state that was

1:35financially bankrupt . At the time , the

1:37debt was around 65 % of the country's

1:39annual wealth . On September 9 , on

1:41France 5 , the editorial director of the

1:42daily newspaper L'Opinion explained

1:44that we had reached the end of the end

1:46of the end of a system . The presenter

1:48reminded him that 20 years ago , he

1:50published a book entitled France in

1:51bankruptcy . The public debt was then

1:53around 1,000 billion euros . Today , it

1:55is 3,536 billion , or 117.5 % of

1:58everything the country produces in a

2:00year . So , was he wrong ? No. He was

2:02right about the trajectory . The debt

2:04has multiplied by 3.5 % . The problem

2:07lies elsewhere : 20 years of warnings

2:08end up making the warnings inaudible .

2:10And an inaudible warning is a gateway

2:12to the opposite promise . If the

2:14disaster never happened , then perhaps

2:16the debt could be wiped out with the

2:18stroke of a pen . Except that something

2:20has changed , and that is the price of

2:22the debt . For 10 years , France has

2:24borrowed at rates close to zero . Today ,

2:27these loans are coming due and need to

2:29be replaced at rates that are currently

2:31around 4 % . The interest bill for the

2:33French government is therefore

2:34increasing on its own without any

2:36political decision being made . Around

2:3860 billion this year , more than 70 next

2:41year , more than 90 in 2029. The

2:43increase in just one year is about as

2:45much as the entire budget of the

2:47Ministry of Justice . So that's why the

2:50idea is finding an audience . We're not

2:54talking about cancelling the debt , to

2:56be precise . We're talking about

2:57cancelling a fraction of it , and it's

2:59important to clarify that . 488 billion

3:01out of 2,900 , approximately 1 / 6th holds

3:04the rest of this debt . Well , in the

3:07first quarter of 2026 , 57.5 % is in the

3:09hands of foreign investors . A little

3:12over 10 % is held by French banks , a

3:14little under 10 % by French insurers . So

3:16when you hear about French debt , you

3:18actually mean Japanese funds , American

3:20funds , banks , insurers who manage your

3:22life insurance and the Bank of France .

3:24Every euro that the State owes is a

3:26euro that it owns and that it relies on

3:28. Obviously , the proposal only affects

3:30the last block . No savers , no insurers ,

3:33no pension funds are affected , and the

3:35cancellation camp is right to repeat

3:37this . The question remains what to do

3:38with this share . Here , the debate

3:40constantly confuses two completely

3:42distinct tasks . To distinguish them , a

3:44single question is enough . Does the

3:46State release the money on the day this

3:48debt has to be repaid ? First answer :

3:50yes , it pays . A debt instrument is a

3:53loan with an end date . France borrows

3:55€ 100 today . It agrees to repay the

3:57€ 100 in 10 years and to pay interest

3:59each year in the meantime . At the end

4:02date , the State makes the repayment ; it

4:04pays out money . Except it doesn't

4:05reduce its debt a cent . Why ? Because

4:08the next day , it has to borrow the same

4:09amount again to balance its budget . A

4:11State never repays its stock of debt .

4:13In reality , it does this by rolling

4:15over . Each loan that ends is replaced

4:18by a new loan . So the real question is

4:20never whether the State can pay it . It

4:22is whether it will find someone to lend

4:24it the money needed to repay the

4:26previous debt . Freezing the debt

4:28answers that question . The Bank of

4:30France is reimbursed and with the money

4:31it has just received , it immediately

4:33buys back a new security . The State has

4:35paid with one hand and borrowed again

4:36with the other . Same day , same lender ,

4:38its cash balance hasn't moved a euro .

4:40And the debt does not move either . It

4:41remains parked at the central bank

4:43indefinitely without ever returning to

4:44the market . Hence the term , it is put

4:46on ice . This is what the dismayed

4:47economists propose . The Treasury repays

4:49the principal to the ECB or the

4:51Eurosystem , which can then reinvest the

4:52sums recovered . They also consider in

4:54the same text transforming these

4:56securities into perpetual debt , that is

4:58to say a loan with no end date that the

5:00State never repays but on which it

5:01continues to pay interest . And frozen

5:03is nothing new . The European Central

5:05Bank did this for nearly 10 years on

5:07the securities it had purchased since

5:082015 and then on those it purchased

5:10during the Covid pandemic . Each time a

5:12security expired , it bought another one

5:14and then it stopped in 2023 for the

5:16first program , in January 2025 for the

5:18pandemic program . The money repaid is

5:20no longer invested and the stock has

5:22shrunk to 546 billion at the end of

5:24December , 488 billion at the end of

5:26June , 58 billion in 6 months . None of

5:28these securities have been sold . They

5:30have simply matured . The State repaid

5:32them , and the Banque de France did not

5:34buy them back . The stock that Jean-Luc

5:36Mélenchon wants to burn is therefore

5:38already diminishing due to a decision

5:39taken in Frankfurt in 2023. And as it

5:41diminishes , someone has to take its

5:43place . Every security that the Bank of

5:45France does not buy back , the Treasury

5:46must sell to someone else , an investor .

5:48The French State's primary creditor has

5:50withdrawn from the market , and it did

5:51not do so by selling , it did so by

5:53stopping buying . The second possible

5:55answer to the question " Does the State

5:56pay ? " Well no , the State never weighs

5:58in . The paper is torn up , the line

6:00disappears from the meter . That is

6:02cancellation . And that is where we

6:03enter into an operation that no major

6:05central bank has carried out in modern

6:07times with its own state's debt . When

6:09Jean-Luc Mélenchon says in the fridge

6:11or in the fire , he isn't choosing

6:12between words . He chooses between a

6:15routine operation and one without

6:16modern precedent : reversible on one

6:18hand , definitive on the other . And this

6:21difference is obviously more than a

6:23difference in vocabulary because it is

6:24exactly the one on which European law

6:26has settled . The treaties prohibit the

6:28central bank from financing states .

6:30Frozen , no one disputes it , the state

6:33repays . It is the second option that

6:35raises questions . Does erasing a

6:36state's debt constitute financing it ?

6:38Yet this point already went before the

6:40courts and is worth a look . In 2015 ,

6:42when the European Central Bank began

6:44buying up hundreds of billions of

6:45dollars in state debt , many German

6:47lawyers and savers took legal action .

6:50Their argument was that the treaties

6:51prohibit the central bank from

6:53financing states , and that is exactly

6:54what it does . In December 2018 , the

6:56Court of Justice of the European Union

6:58ruled against them . It noted that the

7:00central bank does not buy from the

7:02state , it buys on the market and , above

7:04all , that it has imposed safeguards on

7:06itself , ceilings per issue , quality

7:07criteria and a simple rule . No one

7:09should be able to count on it . The

7:12court wrote that a private operator

7:13cannot be certain that its securities

7:15will be bought back and that a state

7:17cannot therefore relax its budget by

7:18relying on the central bank . This is

7:21only on condition that the operation

7:23remains monetary policy . A year and a

7:25half later , the German Constitutional

7:27Court ruled that the European Court had

7:29exceeded its jurisdiction , by failing

7:30to seriously review the program . Two of

7:33the highest courts in Europe

7:34contradicted each other on the same

7:36operation and it has never really been

7:37settled . As for the cancellation ,

7:39however , there is nothing . No judge

7:41said anything since no one has ever

7:43done so , but the implication is obvious

7:45. The European Court validated the

7:47buybacks solely because nothing was

7:48ever acquired by the State . To cancel

7:50is precisely to acquire something

7:52definitively from it . What has already

7:54been decided , however , is the

7:55arithmetic beyond the legal . So let's

8:00do the exercise for real . You are the

8:02Minister of Finance and you decide to

8:04cancel these 488 billion euros . Take

8:06the Banque de France balance sheet ; on

8:08the left is what it owns , and French

8:10securities are there . You cross it out .

8:11The ratio that everyone looks at , the

8:13debt-to-GDP ratio , goes from 117 to

8:15just over 100 % . On paper , it looks

8:17great . Except a balance sheet has two

8:19columns , and you only crossed out one

8:20of them . Opposite , on the right , is

8:22what the Banque de France owes . And

8:24what it owes is thousands of billions

8:26of euros that belong to the banks to

8:28which it pays interest . Every year , you

8:30have just eliminated an income and the

8:31expenditure is still there . So where do

8:33these thousands of billions come from

8:35and why do they have to be paid ? A

8:39central bank has a mandate : to keep

8:42prices stable . And to do this , it has a

8:44main instrument which is to set the

8:45price of money . When it lowers its key

8:48rate , borrowing becomes cheaper for

8:50everyone : households , businesses , the

8:51government , etc. Activity picks up and

8:53that pushes prices up . When it raises

8:56its key rate , the opposite happens and

8:57it tries to combat rising prices . In

8:592014 , the ECB lowered its rate very ,

9:02very low , but it wasn't enough . Europe

9:04was on the verge of deflation with

9:05falling prices , which is what a central

9:07bank really wants to avoid . So it came

9:09up with something else . Starting in

9:11March 2015 , it began buying government

9:13debt on the markets on a massive scale ,

9:16tens of billions per month , then much

9:18more with the pandemic in 2020. For 6

9:20years , every time France borrowed ,

9:22there was a virtually guaranteed buyer

9:24with deep pockets . That's what drove

9:26rates down to zero . And that's why the

9:29Bank of France now has nearly 500

9:30billion euros worth of French

9:31securities in its vaults . Now , what

9:33money did it use to do all this ? Well ,

9:35just like you have an account at your

9:37bank , each bank has an account at the

9:39Bank of France . That's where banks

9:40settle with each other and that's where

9:42the central bank acts . When the ECB

9:44bought its securities , it didn't go and

9:46get the money from a vault . There's no

9:48vault , it took the securities and in

9:49fact , it wrote a larger number on the

9:51account of the bank that sold them .

9:53This number didn't exist the day before

9:54, it does now . These numbers have a

9:56name . They are the reserves , the

9:57deposits that the banks hold at the

9:59Bank of France . Here's your right-hand

10:01column . The two columns were created on

10:02the same day . In the same transaction ,

10:04every euro of debt bought on the left

10:06created € 1 in reserves on the right .

10:08That's why crossing out just one of the

10:10two columns doesn't solve much . The

10:11central bank pays interest on these

10:13reserves . It pays interest to banks on

10:15the money it has created itself . So ,

10:17imagine that your bank pays you 4 % of

10:19your current account balance each year

10:21with money it has credited itself . Why ?

10:24Because it has become the central

10:26instrument of monetary policy . In

10:27reality , the rate at which it pays

10:29interest on its reserves is the key

10:31rate . It is by setting it that it sets

10:33all the other rates in the economy ,

10:34such as your mortgage and your savings

10:36account . This rate was zero or even

10:38negative until July 2022 , and then it

10:40quickly rose to 4 % in September 2023. It

10:43has come back down , and has just risen

10:45to 2.5 % since September 16 , 2026. When

10:48the rate goes up , the bill in the

10:50right-hand column goes up with it . So ,

10:52if we take the figure at the peak of

10:54the increase in 2023 , this remuneration

10:56would have cost around 125 billion

10:58euros per year to all the central banks

11:01in the eurozone . That is twice what the

11:03French government pays in interest on

11:04its entire debt in one year . However ,

11:06when Mélenchon's proposal is

11:07criticized for turning the table to the

11:09benefit of the states , the cancellation

11:10camp responds as follows : " It is

11:11already turning to the benefit of the

11:13banks . " The argument carries weight and

11:15this is what monetary policy

11:16specialists respond . A central bank

11:18does not decree rates . It has no way of

11:20imposing them by law . All it can do is

11:23to ensure that no bank agrees to lend

11:25below the level it targets . Before 2008

11:27, it was very simple . Reserves were

11:29scarce . The central bank distributed

11:31them drop by drop and by controlling

11:33this small quantity , it controlled the

11:35price . Debt buybacks have destroyed

11:36this regulation . With each security

11:38purchased , the central bank has

11:39credited a bank's account . The banks

11:41therefore find themselves with

11:42thousands of billions of euros in

11:44reserves that they never requested ,

11:45which are lying dormant and earning no

11:47income . In this situation , a bank lends

11:49at any rate above zero since it is

11:51always better than nothing . And the

11:53market rate collapses , hence the

11:55solution . The central bank announces

11:57that it will pay 2.5 % on any reserve

12:00left with it . From then on , no bank can

12:02lend below 2.5 since this rate is

12:04acquired without risk and without doing

12:06anything . This is the principle of the

12:08guaranteed price . If the State agrees ,

12:10for example , to buy all the wheat at

12:12€ 200 per ton , well no farmer wants

12:14to sell at € 190 . Removing the

12:16remuneration would therefore mean

12:17losing control of the rates , which does

12:18not change the fact that the money does

12:20indeed go to the banks . Indeed , both

12:21things are true in reality . At the same

12:23time , it is a transfer and it is the

12:25lever of monetary policy . So , two

12:27clarifications on the scale . First ,

12:28this bill has fallen significantly

12:30because rates have come back down and

12:32portfolios are deflating . Excess

12:34liquidity in the eurozone fell to 2,358

12:36billion in the spring of 2026 , or

12:38around 60 billion per year , half the

12:40peak . Second , central banks are losing

12:43money . The Bundesbank , for example ,

12:46lost almost 20 billion in 2024 and 8.6

12:48in 2025 and it expects losses for an

12:51extended period . Which means that it

12:53will not distribute any profits to the

12:54German state . The Banque de France lost

12:567.7 billion in 2024. It did indeed post

12:59profits of 8.1 billion in 2025 , but

13:01look where that profit came from . It

13:03sold 129 tons of gold stored in New

13:04York that failed standards and then

13:06bought back the equivalent . Its

13:08reserves are therefore unchanged . This

13:09gives us an exceptional capital gain of

13:1111 billion . If we remove this

13:12transaction , 2025 will also be a year

13:14of loss for the Banque de France . And

13:16when the Banque de France makes money ,

13:17it simply pays part of it to the State ,

13:19like a company to its shareholder . The

13:21last payment was in 2021 for 2020 , 1.9

13:23billion . Since then , nothing , zero .

13:26There are no more dividends . The

13:30cancellation camp has a fundamental

13:31response to all this and it is serious .

13:33A central bank has no creditors in the

13:35ordinary sense . What it owes are euros .

13:38And € 1 buys you nothing more than

13:40just € 1 . If it cancels a debt that

13:42it holds , no one can come and claim

13:44anything from it in return . Better

13:45still , when the State reimburses its

13:47central bank , the money goes nowhere .

13:48It is destroyed . The debt fades on the

13:51left , the reserve on the right . So

13:53canceling does not create any new

13:55currency . It prevents destroying a

13:57currency that already exists . The

13:59founding column of the cancellation

14:01camp , published in Le Monde in May 2020

14:02by this economist , says it in one

14:04sentence : " No burden is transferred to

14:05anyone since its liabilities are not

14:07due to anyone . " The effect favors this

14:09camp . In 2021 , it was announced that

14:11central banks would lose money and that

14:13the State would lose its dividend . B ,

14:15both happened without the slightest

14:16cancellation and nothing happened . The

14:18Banque de France itself writes that a

14:20central bank is there to combat

14:21inflation , not to make a profit . Add

14:23that 10 years of massive purchases have

14:25not produced inflation in the sense of

14:27rising prices in any case and that

14:28Japan has been living for more than 10

14:30years with a central bank that holds

14:31almost half of its public debt . So , two

14:33responses to this from the

14:34anti-cancellation camp . The first is a

14:36distinction . These losses are losses of

14:38income . The central bank collects less

14:40than it pays out , but it still holds

14:42these securities that are worth

14:43something at their purchase price .

14:45Cancellation means erasing the asset

14:46itself , not just the income . And none

14:47of the episodes mentioned is

14:49cancellation . In all cases , the central

14:51bank holds and can resell on the final

14:53transaction . No one has any data on

14:55that . The second is a blind spot . The

14:56euros owed by the central bank were not

14:58created in a vacuum . They were received

15:00. They were received in exchange for

15:01the security it received . Wiping out

15:03the security means leaving the most

15:05liquid currency in circulation without

15:07the debt that faced it . Hence the

15:09statement by Nicolas Dufren , who heads

15:10the Rousseau Institute and has been

15:11promoting this proposal since 2020.

15:13Subsidizing banks is no problem .

15:14Relieving governments is impossible .

15:16Find the mistake . So , everything is in

15:17the word subsidize . A subsidy is a

15:19payment that we decide to make , that we

15:20can stop , and whose stopping leaves the

15:22rest in place . See if it fits here .

15:24Well , these interests are not set

15:25according to anyone , especially not the

15:27banks . They are set according to

15:29inflation . When prices soar , the

15:30central bank raises the rate to make

15:32loans more expensive and cool the

15:34economy . When it falls , it lowers it .

15:35The banks collect more or less

15:37depending on what prices do and at no

15:39time does anyone decide to pay them

15:41anything . Second test . What happens if

15:43we stop ? You don't save 60 billion by

15:44leaving everything else as it is . You

15:46lose the key rate , that is to say the

15:48only link that connects the central

15:49bank to the rate of your mortgage and

15:51that of your savings account . The

15:53payment and the instrument , once again ,

15:54are one and the same thing . It is by

15:56paying this rate that the central bank

15:58imposes it on the rest of the economy .

16:00If you kill this rate , you no longer

16:02have a transmission belt from the key

16:03rate to the economy . Another test , and

16:05it is the one that decides , a subsidy

16:07gives someone income that they would

16:08not have had otherwise . So , ask

16:09yourself what these banks would have

16:10received if the central bank had never

16:12bought these securities . The securities

16:13would have remained in private

16:14portfolios and the interest would have

16:16gone to those who hold them , including

16:17the banks that already held them . The

16:19remuneration of reserves replaces an

16:20income that already existed before it .

16:22It does not invent a new one . The fact

16:24the formula relies on remains , and it

16:26is correct . Tens of billions go to the

16:28banking sector every year without any

16:30parliament having ever debated it , when

16:32a transfer of the same order to the

16:33State triggers a nervous breakdown . The

16:35word subsidized does not describe the

16:38mechanism well , but the symmetry is

16:40still real . Let's take stock , we've

16:41moved ahead quite quickly . This entire

16:43discussion concerns 1 / 6th of the French

16:45debt , the portion held by the Bank of

16:47France . Behind the slogan , there are

16:49two very distinct operations . Freezing

16:50what is already being done in reality

16:52without anyone finding fault with it ,

16:54or erasing what no one has ever done .

16:55The erasing problem doesn't come from a

16:57creditor coming to claim what is due ;

16:59it comes from the balance sheet .

17:00Opposite these securities , there are

17:01the banks ' reserves , they remain there

17:03no matter what happens and they cost

17:04because they are remunerated at the key

17:06rate and tearing up the security does

17:07not make that invoice disappear . So

17:09let's go back to your sheet with the

17:10right-hand column at the top . You've

17:11crossed out the left , but you still

17:12have to pay for the right . So with what

17:14? This is where two economists , Olivier

17:19Blanchard , former chief economist at

17:21the International Monetary Fund , and

17:23Quentin Van de Veyer , professor of

17:24finance at the London Business School ,

17:26posed the problem most clearly in the

17:27world on September 4 , 2026. The

17:29supporters of cancellation , they write ,

17:31promise three things at the same time .

17:33Real budget margins , no casualties , and

17:35controlled inflation . And unfortunately

17:37, arithmetic only allows for two . Three

17:39doors , you're still a minister and

17:40you're going to open them one by one .

17:42Each time , you'll see that one of the

17:44three promises falls through . First

17:45door , you're the one who pays . The

17:46State fills the hole that has just been

17:48dug in the central bank , either by

17:49recapitalizing it all at once , or by

17:50paying it each year , well , the

17:51equivalent of the interest has

17:53disappeared . Its stated debt is down ,

17:54but the State fills the gap . It now

17:56pays the interest it no longer pays to

17:58the Bank of France so that it can pay

17:59the banks in turn . Its deficit has

18:01therefore not changed . And if it

18:02suddenly recapitalizes , it is worse in

18:04the short term because a

18:04recapitalization is counted as a

18:06deficit in the year in which it is made

18:07. So , it will be objected that the

18:09cancellation costs nothing since the

18:10Bank of France pays its profits to the

18:11State . It pays interest , and we give it

18:13back to it in dividends . That is

18:14exactly what it implies . If interest

18:16and dividends cancel out , then

18:18eliminating the interest also

18:19eliminates the dividend . The profit

18:20melts away . And this dividend has not

18:22disappeared since 2021. It is simply

18:24deferred . The losses remain on the

18:26balance sheet to be offset later by

18:27future profits . Al , you have just seen

18:29what it took for a single year in the

18:31glass . Nicolas Dufren answered all this

18:33on September 11 , 2026. What matters , he

18:35writes , is the repayment of the

18:37principal . Each security that matures

18:40is fully repaid to the holder . What the

18:42State never repays is the stock . The

18:44matured security is replaced by a new

18:47one . The only question that remains is

18:49who buys the new security when the Bank

18:51of France no longer does so . The fridge

18:54answers that one fully without

18:55canceling a thing . He says something

18:56else himself , a bit more awkward for

18:58his thesis . He has argued from the

18:59outset that cancellation should be

19:01conditional on reinvestment and he

19:03specifies , in parentheses , what that

19:04implies . With a reprint , we erase

19:06nearly 500 billion and we go back to

19:08borrowing to finance what the erasure

19:10was supposed to allow . Let's return to

19:12this first door . The State erases what

19:14it owes to its central bank , then it

19:15pays it enough each year to pay the

19:17banks . The line changes name , but the

19:18amount does not change . The dividend

19:20that is presented as the gain from the

19:21operation is the same money counted a

19:23second time . And the only question that

19:25the cancellation claimed to resolve ,

19:26finding a buyer for the new securities ,

19:28well the freezing already answers that

19:29without erasing anything . At first ,

19:31what is the verdict ? Indeed , no victims

19:33, no inflation but also no margin . The

19:36first promise falls . Second possible ,

19:38no one pays . The central bank manages .

19:40At this stage , it has two paths . The

19:41first , it stops remunerating the

19:43reserves . No more interest to pay , the

19:44problem is solved . But that rate , as

19:46said , is its steering wheel . And

19:47releasing it poses an immediate

19:49question . What can slow down an

19:51overheating economy ? So , we will

19:52respond that a central bank has managed

19:54itself differently and it is true that

19:56before 2008 , neither the ECB nor the US

19:58Federal Reserve paid for excess

20:00reserves . They kept them scarce and set

20:02their rate through their daily

20:03operations . Except that we are not

20:05going back to scarce reserves with

20:062,358 billion in surpluses . To make

20:08them scarce , we have to immobilize them

20:09. And immobilizing reserves without

20:11paying for them is already the next

20:12step and we will see it right after .

20:14But first , the second path , the money

20:16created to pay the interest is paid

20:17into the banks ' accounts , that is to

20:19say into the reserves themselves . The

20:20right-hand column therefore increases

20:22by the amount of the interest and the

20:23following year , interest must be paid

20:25on a larger amount . The bill feeds

20:26itself at the rate of the key rate .

20:29Blanchard and Van Deveilleur have an

20:30image for this . Fighting inflation by

20:32creating money is like bailing out a

20:34boat while piercing the hull . So ,

20:35second door , verdict , clearance . Yes ,

20:37no identifiable victim at the time , but

20:40inflation . The third promise falls . And

20:42inflation is a victim all the same ,

20:44simply spreading to all those who hold

20:46euros . Third door , the banks pay . They

20:48are forced to leave part of their money

20:50at the central bank without being paid

20:52for it . Unpaid mandatory reserves . The

20:55interest to be paid decreases

20:56accordingly and this time real margins

20:57appear . A precision that gives a bit of

20:59scale . Since September 20 , 2023 , the

21:01ECB no longer pays any remuneration for

21:03mandatory reserves . It has already

21:05opened this door a crack on its own

21:07initiative and the mandatory reserve

21:10ratio has been 1 % since 2012 , or 172

21:12billion euros . To immobilize nearly 500

21:14more , it would have to be raised to

21:16nearly 4 % . What about tripling this

21:18mandatory reserve ratio ? How much does

21:19that bring in ? Well , 488 billion that

21:21we stop paying at today's rate , that's

21:23about 12 billion euros per year . The

21:26government seeks 30 billion in 2027

21:27budget savings ; that is convenient , as

21:29it covers more than a third of the

21:30total . Except that when talking about

21:32annual bills , we can perhaps reassure

21:33ourselves at a good price . Quentin

21:35Vanilleer pushed the calculation

21:36further . If we freeze our reserves at

21:38zero , it's not for 1 year , it's forever

21:40. In the very logic of cancellation ,

21:41the money created has become permanent

21:43and the repayments that would have made

21:45these reserves disappear will never

21:47take place . However , an investment that

21:48is blocked forever and earns nothing is

21:50worthless . Anything cheaper is

21:51worthless . So it wouldn't be 12 billion

21:53per year to absorb . It would suddenly

21:55be the equivalent of what the equity of

21:56the major French banks is worth today .

21:58That point is not guaranteed , it must

22:00be said . The accounts would have to

22:02actually recognize this loss and the

22:03freeze would have to be definitive . But

22:05that is by far the most likely outcome

22:07because that is exactly what the

22:08operation does . So , third door , verdict

22:11, closing . Yes , no inflation , yes , but

22:13a victim . The second promise falls .

22:15Three doors , three promises , each time ,

22:17one is missing . That is the trilemma

22:19that faces us when we want to cancel or

22:20freeze the debt . And now , take your

22:25sheet of paper one last time because

22:26there is something that this debate

22:28never says . You are still a minister

22:29and you have on your desk the 12

22:30billion from the third door . That is

22:32what no one will tell you . The third

22:34door , the only profitable one , needs no

22:36cancellation . You can immobilize the

22:37same amount of reserves without paying

22:39them without touching a single security

22:40. The Bank of France stops paying the

22:42same interest . It collects what the

22:44State pays , and the gain goes back to

22:46the budget as a dividend . Same money ,

22:47down to the euro . What the cancellation

22:49brings in addition is therefore

22:50absolutely not money , it is a figure .

22:52The aid to GDP ratio goes from 117 to a

22:54little over 100. The operation is made

22:56up of two parts that are actually

22:58presented as one . The part that brings

23:00in money is the measure on reserves .

23:02And the part that is visible is the

23:04cancellation . And it is the second that

23:05carries all the public debate while the

23:07first sends the bill . Now we must

23:09decide who gets the bill . Because a

23:13bank that is forced to pay money that

23:15is dormant is not going to absorb the

23:17loss silently . It makes up for it

23:18through the remuneration of your

23:20savings , the interest rate on your

23:21credit or what it pays to its

23:22shareholders . And on that last word ,

23:24beware of the image it conveys . Hey ,

23:26the shareholders of French banks are

23:27not just anonymous funds . The three

23:29major mutual insurance groups , for

23:31example , belong to 31 million members ,

23:33i.e. their own clients . Employees own a

23:35share of their bank and the savings of

23:36French people with life insurance

23:38policies own another . So when we say

23:40banks pay , we are also talking about

23:41these people . It even has a second exit

23:43and it is mechanical . Its fixed

23:45reserves are calculated in proportion

23:46to what the bank owes its clients ,

23:47deposits and short-term securities with

23:49maturities of less than 2 years .

23:50However , a bank does not collect these

23:52liabilities , it creates them . When it

23:53grants a loan , it does not seek money

23:55from another client . It credits the

23:57borrower's account and creates the

23:59deposit then . The direction of the

24:00mechanism is the opposite of what is

24:01often imagined . It is not deposits that

24:03create loans , it is loans that create

24:05deposits . And to have fewer frozen

24:07reserves , a bank must reduce its

24:08balance sheet , which means granting

24:10less credit . There is one loophole left

24:12, and it is in the calculation . Assets

24:13over two years are excluded from the

24:15tax base . So a bank can fund itself

24:17longer instead of lending less . Except

24:19longer-term financing costs more . In

24:21both cases , someone pays . Now this

24:23mechanism has a name . It is called

24:24financial repression . A levy on those

24:26who hold money , who do not go through

24:28any vote and who do not appear on any

24:29tax forms . Now look at who is proposing

24:31this third way . It is still Nicolas

24:33Dufren in his note Doubt when he comes

24:35to the objection of reserves , here is

24:36his answer . We are not obliged to

24:38remunerate them . The same monetary

24:39policy orientation would be obtained by

24:42sterilizing excess liquidity by means

24:44of unremunerated mandatory reserves . A

24:46technical method presented as neutral .

24:48But weeks later , Blanchard and Van de

24:50Veilleur described the same operation

24:52and said that who pays for this

24:53operation , which economists call

24:55financial repression , is a tax that

24:56does not say its name , ultimately borne

24:58by bank depositors , their shareholders

25:00and their borrowers . On September 11 ,

25:02Dufren replied , taking up the trilemma

25:04and accepting it . The operation can

25:06therefore generate real margins with

25:08controlled inflation but not without

25:09victims . However , we will see that the

25:11designated victim , the banks , is not

25:12really one . Put this sentence next to

25:14what he wrote 15 days earlier in his

25:16note . Word for word , the operation has

25:18no victims . Further in his reply , he

25:19follows that logic through . If this

25:21sacrifice , he writes , allows the

25:23community to avoid hundreds of billions

25:25in repayments and to devote its

25:26resources to investment , then it seems

25:28to him to be largely justified . The

25:30word is his . Sacrifice . In 15 days , we

25:32went from a victimless operation to a

25:34justified sacrifice . So , this is a

25:36position that is perfectly defensible ,

25:38or at least arguable , but it is no

25:39longer the same . The first said , " Money

25:41makes it possible to erase a debt

25:43without anyone paying anything . " The

25:44second said , " Someone pays and it's

25:46worth it . " The first is a monetary

25:48argument . The second is an ordinary

25:50political arbitration of the kind that

25:51we prefer to settle by a vote normally .

25:54Al remains the word because Blanchard

25:55and Vant de Veilleur's sentence does

25:57not mention banks , it mentions bank

25:59depositors , their shareholders and

26:00their borrowers . The question therefore

26:02becomes very concrete . When a blow is

26:04imposed on a bank , does it stay with

26:06the bank ? Two elements of response . The

26:08first , which the French debate has left

26:09aside , is monetary policy , it is common

26:11, it is not French . Mandatory reserves

26:13are not decided country by country .

26:15This applies to all banks in the

26:16eurozone in proportion to the deposits

26:17they collect . So French debt would be

26:20cancelled and the bulk of the bill

26:21would be paid by German , Italian ,

26:23Spanish , Dutch and other depositors . So

26:25this does not prohibit the operation .

26:26This explains why it is not really

26:28decided in Paris . The second is

26:29empirical . There are two academic

26:31studies on what happens when a bank's

26:32balance sheet is taxed and they tend to

26:34agree . The study of the 2010 Hungarian

26:36tax published in the Finance Review

26:38concludes that the tax was passed on to

26:40the least mobile customers , households ,

26:42and all the more so when the market is

26:43concentrated . So think about it . The

26:46French retail banking market consists

26:48of only six groups . The second study

26:50published in the International Tax and

26:51Public Finance extended to a panel of

26:53countries and specifies where the bill

26:55goes . The banks mainly pass on the tax

26:57to borrowers by increasing the interest

26:59rates on loans . Note the nuance , the

27:01same study notes that the remuneration

27:03of deposits can increase . So it is not

27:05your savings account that pays first ,

27:07it is your credit . And it is possible

27:09that the banks absorb everything from

27:10their profits as Nicolas Dufren writes .

27:12But this remains to be proven and the

27:14only two studies available today say

27:16the opposite . A country can perfectly

27:18well decide democratically that its

27:20savings finance its state . France did

27:21this for 30 years after the war and

27:23Dufren cites this precedent . The fact

27:25is entirely accurate . This precedent

27:27was called the treasury circuit and it

27:29consisted of forcing French banks and

27:30savers to finance the state . Both camps

27:32cite the same 30 years and describe the

27:35same thing . There remains one objection

27:39that does not come from the opposing

27:41camp , it comes from an economist who

27:42defends freezing . On the France 5 set

27:44on September 9 , Clara Léonard ,

27:46co-founder of the Avant-Garde Institute

27:48, pointed out that the proposal

27:49concerns non-problematic debt , which is

27:51the debt held by the European Central

27:53Bank . And she added that if we freeze

27:55this debt , well , all that remains is

27:57held by private investors . So , take the

27:59sheet again , the securities held at the

28:01Bank of France are in fact the only

28:03fraction of French debt that is already

28:05outside the market . No one sells them

28:06one morning because a poll has changed .

28:08No one demands risk premiums on them .

28:10That is the quiet part and that is

28:12exactly the one we are proposing to

28:13erase . The other five , 6th , are on the

28:15market and France returns there every

28:17week , in the order of 320 billion euros

28:19to be invested in 2026 in the medium

28:21and long term , of which 174 billion is

28:23just to replace the maturing securities

28:25. Look at the issue from the

28:27other

28:27end . You are no longer a minister , you

28:29are the fund that holds part of the

28:30remaining 5e and you have just seen a

28:32state wiped out with a stroke of a pen ,

28:34the only debt over which it had control

28:35. What do you deduce from this about

28:37yours ? And if you think the market

28:38would take it without flinching , hear

28:40the reassuring argument . It is said

28:42that at each auction , demand greatly

28:44exceeds supply , sometimes two and a

28:45half times . On the same panel , Anne

28:47Lord Kichel , who heads the Global

28:48Sovereign Advisory firm and advises

28:50States on their debt , dismantled the

28:51argument . Most of those who buy the

28:53French head are funds that replicate

28:55indices . They want to gain exposure to

28:57the eurozone and France is in the index

28:58. Even without liking France , in fact ,

29:00people buy it . It is a mechanical

29:01demand and these are the words . It can

29:03also disappear quite quickly . This risk

29:05does not appear in any of the three

29:07doors . It is not at all

29:08accounting-based , it is behavioral .

29:13Which brings us to the question we

29:14haven't asked . We have spent some time

29:17on 1 / 6th of French debt . What holds the

29:19other five together ? What makes a

29:21Japanese fund or a French insurer lend

29:23to France at 4 % rather than 9 % ? No

29:25assets , the State has not put anything

29:27up as collateral , and neither has the

29:28promise to repay a stock , since this

29:30stock is never repaid , it is refinanced

29:31. What the lender really relies on is

29:34not being repaid . Each security is

29:35repaid when it becomes due , it is that

29:37someone will still be able to lend one

29:38day when the security needs to be

29:40replaced . And behind this ability to

29:42find a lender , there is a huge

29:44hypothesis that no one in Lork

29:45discusses , which is consent to taxation

29:47. If you read the rating agency reports

29:50, she says , this is one of the reasons

29:52why France still has such a high rating

29:53. These rating agencies count it as a

29:56strength . What guarantees French debt

29:58is therefore the willingness of the

29:59French to let themselves be charged . It

30:01is a behavior , not an asset , not a

30:03reserve , not a pledge . And according to

30:05her , a crisis of this consent would be

30:07much more serious than having a debt of

30:09100 or 125 points . So during this

30:11entire discussion , we fought over

30:12balance sheet lines : a sixth of the

30:14debt , a right-hand column , a ratio that

30:16goes from 117 to a little over 100.

30:17Greece is a reminder of what these

30:19lines are made of the day the lender no

30:20longer shows up . Two scanners for all

30:22of Athens , one math teacher for three

30:24high schools and 15 years later ,

30:25production that still hasn't returned

30:27to its level . So this doesn't prove

30:29either side right ; it shifts the issue .

30:31What decides the fate of an indebted

30:32country is never the level reached by

30:34the meter . It is the moment when there

30:35is no one left to lend . One last point ,

30:37and it comes from Nicolas Dufren

30:39himself . On September 9 , Quentin Van

30:41Deilleer criticized him for wanting to

30:43finance the State with money rather

30:44than debt . His response was not a

30:46denial . Yes , he wrote , money is now

30:48becoming permanent and intentional . The

30:50ban on monetary financing was

30:51established as dogma 50 years ago and ,

30:53according to him , it deserves a

30:55democratic debate . And a ban is not an

30:57economic argument . The question is

31:00clearly posed and it is political . A

31:02monetary rule remains a decision and a

31:03decision can be discussed again . So we

31:08come back to the question that was

31:10raised by this speech by Jean-Luc

31:11Mélenchon at the end of June 2026. Can

31:13we burn France's debt ? Well technically

31:15yes , there is nothing materially

31:16preventing it . These securities are not

31:18a claim on anyone outside . The State

31:20owes money to itself through an

31:22institution that it owns . It would

31:23require a legal battle , a power

31:25struggle with our neighbors and someone

31:26to collect the bill , but it is not

31:28impossible . The 12 billion from the

31:29third door , we can have them by

31:31tomorrow morning , intact . Except that

31:33we have to stop for a second on this

31:34answer because it should make us all

31:36tick . Take any debt in your life , your

31:38mortgage , the money you owe to a loved

31:40one , a bill from a craftsman and so on .

31:42None of them can be erased because you

31:43have decided that they should be erased

31:45. There is always someone on the other

31:46side , and they have a say too . Not here

31:48. 488 billion euros could go up in

31:50smoke because the debtor and the

31:52creditor are in reality the same person

31:54and because the unit in which all this

31:56is counted is manufactured by one of

31:57the two . So , none of this is hidden ,

31:59right ? It's written in treaties , argued

32:01in court , defended on 8pm talk shows ,

32:03and so on . But ask yourself the

32:04question honestly : what does a system

32:06look like where the question of

32:07cancellation can arise ? A modern

32:09currency is based on a pile of promises

32:10. The promise of the State to repay its

32:13lenders , the promise of the Central

32:14Bank to maintain prices , and underneath

32:16, the promise of the French people to

32:17continue paying taxes . There's nothing

32:19else underneath that . Not a gram of

32:21metal , not a pledge , not an asset , not

32:22land , nothing . And a promise , by design

32:24, can be renegotiated . In the meantime ,

32:26the meter is ticking . Interest on the

32:28debt has become the second largest item

32:30in the State budget , at 58 billion

32:31euros in 2026 , behind education and

32:32ahead of everything else . Whether or

32:37not the debt is being burned , the very

32:39existence of this debate has the merit

32:40of showing the disconnection from

32:42reality of the monetary system . A

32:44monetary system that erodes the value

32:46of savings to finance the

32:47irresponsibility of public authorities .

32:49Since 2008 , there has been an

32:51alternative in Bitcoin , a rare currency

32:53that is immune to political

32:54manipulation . If you want to start

32:56saving in Bitcoin in a simple and

32:58secure way , you can find a promo code

32:59in the description to join Bitstack . If

33:02you enjoyed this video , feel free to

33:03like it . We publish a video every week

33:05on the channel . So , subscribe so you

33:07don't miss them , it helps us a lot .

33:09This was Alexandre for Bitstack and

33:10I'll see you next week .