Damien Soitout
Behind the Ledger explains how money is created off the public banking screen, from years of work on financial networks.
A ledger a person can read. The record that settles large money sits on servers behind it.
I spent years working with intelligence agencies whose focus was financial networks. The job was not to watch a person spend on a card. The job was to see where value is created, where it is allowed to move, and who can stop it. What the public calls the banking system is a window. The work sits behind the window. Behind the Ledger is that account. I published it on 1 June 2025, in English, as the third book of A Quick Guide. It runs 153 pages. If you want the object itself, it is here: Behind the Ledger.
Most people meet money as a number in an app, a note in a wallet, or a line on a statement. That meeting is real. You can pay for dinner with it. You can be refused by it. It is also the surface. Between 1998 and 2000, while the public argument was about the year 2000 and the first internet boom, the institutions that actually move large money split the system into two layers that still depend on each other. One layer is what you can see. The other is where the large settlement happens. I wrote the book because those two layers get taught as one thing, and they are not.
The layer you can see is the on-ledger network. It is the retail face: cash machines, card payments, the phone app, the website of the bank, and the ordinary transfer from one person to another. It is built, in the main, on older Windows systems. That is one reason it breaks, and one reason it is full of checks. It is supposed to be visible. It is supposed to fit inside the rules a customer can be shown. It is also fenced. A personal transfer often stops somewhere between ten thousand and two hundred and fifty thousand dollars, depending on the country, not because the wire is tired, but because a larger sum is not meant to live in that sandbox. When a company tries to push a very large payment through the window, the delay is the time required to touch the other layer.
Even a transfer between banks, the kind an officer calls a bank-to-bank movement, is often a change of numbers on that visible ledger. The entry updates. The value, the part that matters when the sum is large, settles off to the side. The clearing days people treat as a technical pause are the time that second settlement takes. If you analyze an economy using only the numbers from the window, you are reading the shadow on the wall.
The other layer is the off-ledger network. People in the work also call it the offline network. It does not run on the customer app. It runs on hardened Unix machines, mostly Linux, over encrypted channels that do not travel the rails a card payment uses. Two movements matter. Server to server is the continuous settlement between institutional nodes, a handshake that reconciles what is owed. Server to bank is the pipe that lets an authorized institution put liquidity in, or take it out, without dropping that flow onto a consumer screen. Access is not a login you apply for with a passport photo. It sits with central-bank units that hold a Federal Reserve authorization, with treasuries that hold a clearance the European Central Bank or the International Monetary Fund will recognize, and with prime banks that meet the Basel standard those rooms actually use.
This is where monetary base is originated outside the story of a printing press, where a cross-border project can be settled in one movement, and where emergency liquidity appears in a crisis. The same design keeps the public ledger, the public audit, and the public disclosure off that table. I am not decorating a thriller. Proponents of the design call the secrecy the price of keeping large flows stable. The price is also that a citizen cannot see the lever.
At the center of the off-ledger layer is the Farm. A server farm, in the ordinary sense, is a room of machines. In this architecture the Farm is the network that sits over that. The banking servers, and in the end the internet servers that carry them, are mapped under its supervision. It is run by the private side. It does not belong to a government, a party, a commonwealth, or a public agency. It is not an ICANN address plan. It is the place that decides whether a unit of money is valid before that unit ever becomes a balance a customer can see.
The Farm is divided into 256 sectors, numbered from 0 to 255. Sector 0 is the nucleus. A handful of principal moderators can enter it. It is the root of trust. The other sectors are spread across physical and virtual nodes, and they handle both the on-ledger and the off-ledger work under authority delegated from that root. Every internet-connected server is, in this design, a participant, whether or not the owner of the machine has ever heard the name. No state owns it. Since 2015 the Federal Reserve has been the liaison between the Farm's authorities and the institutions downstream. Liaison is not ownership. The Fed can carry the message. It does not sit above the protocol. You can read what the Fed says it is, in public, on the Federal Reserve's own page. The book is about the layer that page does not draw.
Farms 42 through 107 are the ones that carry the offline life of the banking system. Farm 42 is easy to misunderstand. It does not hold the money. It holds the archive of how the money was generated and moved, the record of creation and transfer for the farms that follow it. The money itself sits from Farm 43 onward. By Farm 92 you are in funds whose origin is illicit. The ladder between those rooms is the point of the third chapter, and I will put it in plain words below. The mining of the raw data, the thing that later becomes a currency unit, is a procedure inside that secured environment. It is not a press and it is not a gold window. It is data that becomes negotiable only after it has been validated.
Raw money data is the first form. It is a cryptographic pattern on a secured server. It is not yet a euro, a dollar, or a pound. It becomes server money only after the system has taken its costs and its taxes and the remaining unit has been pegged to a national currency. More than nine tenths of that creation, in the flows I worked, is pegged to the euro, because the banks prefer it and because the euro sits deep in cross-border settlement. Dollars and sterling are mined too. They are not the bulk. That is why the European Central Bank is one of the first doors a creation or a transfer of off-ledger funds has to pass. The ECB's public role, monetary policy for the euro area and the right to authorize euro banknotes, is described on the ECB's site. In the off-ledger path, that institution is also the office that keeps the euro's balance from being invented without a signature.
The ratio I was taught, and that I put in the book, is severe. Seven eighths of the raw unit is consumed to produce one eighth of offline currency. That remaining eighth, after legal costs and tax, is what can become clean printed money. People hear “money creation” and picture a benefit that arrives whole. The creation I am describing is mostly cost. The creator, not the note, is the power in the room. The fear, inside the work, is of the creator.
Once the unit is server money it sits in the Farm's treasury until it is sent. The sending uses protected methods. IP to IP, IP to ID, and DTC transactions are the ones that move value from that treasury, or from a sender's server, to the receiver's server. DTC here is the Depository Trust channel used to move what has already been validated, not a retail button. A sender and a receiver both need a license from the monetary authorities before that hop is allowed. Without the license, the data does not become a balance.
Money in this network is not one substance with one moral status. It is classed. On the public side the classes are named M0 through M6. On the off-ledger side the same ladder is NS0 through NS6, and the servers that hold them are MS for the online side and NS for the offline side. The names match on purpose. A clean unit and a poisoned unit can both be called money by a person who has only seen a balance.
M0 and NS0 are completely clean, fully documented, and clear. They can be structured into the boxes that high-value protocols will accept, including IP to IP, IP to ID, and an MT103 manual download. M1 and NS1 are still clean and documented, and they are the funds used for credit, bonds, and secured loans. They cost more to monetize because the file is heavier. M2 and NS2 are legally sound and incomplete. The usual repair is a mirror, a temporary bubble that lets the process move, which is then discarded once the original has been checked. M3 and NS3 are the border. Documents are missing, the procedure was not followed, and a professional thief looks here first, because the file can still be pushed through if someone is willing to rebuild it. M4 and NS4 have no documents. Turning them into currency is laundering, and the amateur thief likes them because the lock is weaker. M5 and NS5 are funds with no origin, or funds already blocked, often tied to serious crime. Monetizing them is possible on paper and absurd in cost, and most jurisdictions freeze them on sight. M6 and NS6 are illegal data, stolen in the middle of a conversion and forced onto the grid. Using them as money costs more than they are worth. They are used to damage a system, not to buy a building.
Only the top of that ladder, M0 and NS0, is allowed into the large protocols. After tax and fees, NS0 and NS1 can cross onto the on-ledger network. Once they sit in a beneficiary account they look like any other balance. They can pay a card, a cheque, or a person. The public then meets M0 and M1 as ordinary money. The meeting is true. The history of the unit is not on the receipt.
Dirty funds, in the narrow sense used in the work, include units broken by hacking, by a system error, or by a fraudulent copy, the bubble that was never discarded. Those units are inert. They cannot be turned into legitimate currency. The board and the inspectorate of the Farm are the people who administer the M0 and NS0 stock. That is a small group, and the smallness is the control.
A transfer from the treasury or from a sender's line to a receiver's line is not a private handshake between two companies. The sender and the receiver need authorization, and the offices that give it are public enough to name. The Federal Reserve is the liaison and the supervisor of the offline network on the dollar side. The European Central Bank authorizes creation and transfer when the unit is euro, which is most of the time, and it runs TARGET2 for the payments the public system will later recognize. Euroclear, in Brussels, enters when offline funds are turned into securities, bonds, shares, or fund units. Its public description of that settlement role is on Euroclear. The International Monetary Fund enters when the same funds are turned into shares, loans, or a guarantee, because those forms touch the balance of payments the Fund exists to watch. The treasury and the central bank of the receiving country then authorize the executing bank. That bank, after it has looked at the sender and the receiver and believes the receiver can hold what is coming, is the last signature before an off-ledger facility is opened. Anti-money-laundering units and, where the file demands it, a court, sit on that path as well.
Every central bank, every prime bank, the subsidiaries, the government lenders, and the private houses are connected to the offline network by servers. The online network is built on top of that connection. The app in your pocket is a client of a system it cannot see. I say that as a person who sat with the people whose job was to see it.
A reader who has followed me for the public questions, how much money exists, when you may spend, what a fixed pile of hours is worth, is asking about the surface in good faith. How much money is there in the world is still the right way to feel the scale of the visible stock. When you can start spending is still a personal question. What you would do with 86,400 is still a way to notice that a sum is also a pile of hours. None of those questions reaches the room where a unit is born. Behind the Ledger is the third quick guide because the first two teach tools you can use in a week, and this one teaches the pipes those tools never touch.
If you remember one distinction, remember this. The on-ledger network is the window: cards, apps, ordinary transfers, and the balances a family actually spends. The off-ledger network is the settlement and the creation, on servers, classed from clean to illegal, and gated by the Fed, the ECB, Euroclear, the IMF, and the bank at the end of the line. The Farm is the private network that validates the unit before either of those rooms will treat it as money. I wrote it down because the window is taught as the whole house. It is the front door. The book is the plan of the rest.
