Money is Created from Nothing (V2)

I want to defend and explain the idea that money is created from nothing.

The idea sounds spooky and paradoxical, and some have used the word ‘magical’ to denigrate the thought, but I’ll explain that properly understood it is none of these things.

The classical definition of 'Money'

Money is classically defined as anything which performs these three functions:

  • Medium of exchange.
  • Unit of account.
  • Store of value.

I won't be questioning this classical definition. My argument – entirely unoriginal, but still surprising to some – is that what plays these three functions is, it turns out, something that can be created from nothing. Before arguing for that, I want to explain in more detail what these three functions mean.

A medium of exchange is whatever is used generally to pay for things. For a society in which there is something playing this role, anyone wanting to make a trade is likely to accept that something in return for whatever they are trading. A society without a medium of exhange would be one which relies only on barter. In a barter society, there no general medium of exchange, what's exchanged when things are traded has to be agreed separately for every transaction.

A unit of account is a scale that is used to express the value of things. This is convenient for many purposes, especially if any kind of trade or lending occurs. For example suppose someone borrows a sheep. What do they owe to their creditor? Well, a sheep. But the borrower might eat the sheep, and perhaps doesn't have other sheep available. So then what does he owe? Clearly something that the borrower and lender agree is equivalent in value to a sheep. Perhaps the borrower makes combs, so the borrower and lender might come to an agreement that the borrower owes four combs. This might be fine for a one-off transaction, but things can get messy fast. If the lender lends to several people, and the borrower borrows from several people, and the lender also borrows, and the borrower also lends, for everyone involved it's going to be hard to keep track if debts are denominated in terms of different units. Combs for debts incurred by the comb maker (and then only when the lender actually wants some combs), sheep for debts incurred by the sheep farmer (when the lender wants sheep), etc. Far more convenient would be for there to be a general way of expressing the value of things. Some unit in which the value of combs and sheep and any other tradable or loanable goods can be expressed. That unit is a unit of account. Logically speaking, just what is chosen as a unit of account is arbitrary. It could for example be combs.

If combs were the unit chosen, although all debts would be expressed in terms of number of combs, debts would not necessarily be paid off in combs. A unit of account is not necessarily a medium of exchange. In fact unit of account is more fundamental than medium of exchange in the sense that societies can, and have, operated with a unit of account but with no a medium of exchange, but not vice versa. If you have a medium of exchange you have to have a unit of account. But a society can have a unit of account without a corresponding medium of exchange. Things could be compared in value according to a common scale, but transactions would involve swaps of whatever was available at the time.

Historically it seems, in fact, there have been societies in which there is an operative way of expressing value but no corresponding medium of exchange. An example is early medieval Ireland: your debt would be measured in terms of a number of slaves, but would be paid by cows, silver cups and broaches; and this was the case well after the institution of slavery had ceased to exist, so no-one in fact could pay their debt in slaves even if they wanted to (Graeber, 2014 p 171).

Although it is possible for a unit of account to not be the medium of exchange, it is not exactly an accident that, for any society in which there is both a commonly agreed unit of account and a widely used medium of exchange, what you use to pay for things is used also to price them. It would be inconvenient to have a different scale as a unit of account from that provided by the medium of exchange. For instance it would be pointless and inconvenient to always state prices in dollars (or slaves or combs, for that matter) but only give and receive payment in pounds.

The final concept from the classical definition is store of value. This means that anything that plays the role of money has to have a value that is reasonably stable over time. This seems to me to be a practical consequence of being a medium of exchange. If the value of someting was not fairly constant over time, no-one would be willing to accept it as part of a transaction.

As I said I am not going to quesion that these functions are definitive. I accept that anything which plays these roles is thereby money. What I want to describe here are some interesting properties of what, in fact, in our present society, plays these roles.

The Basic Concepts.

Now I'm going to introduce some basic concepts in terms of which the nature of money creation can be explained.

 Assets and Liabilities

To understand the creation of money, you have to understand assets and liabilities. Here are a couple of key sentences which introduce and relate these concepts:

An asset, in the relevant sense, is something that is owned and which has a measurable value. A liability a certain kind of obligation with regard to that asset.

All this needs further elucidation and elaboration. I need to explain what I mean by ‘measurable value’, and then describe the kinds of obligation involved. But I shall start with ownership.

Ownership

Although it might seem an obvious concept, the meaning of ‘ownership’ can be difficult to spell out in detail. But we don’t have to go too deep and we can say that, broadly speaking, ownership is a collection of rights to use and to benefit from the thing owned.

Measurable Value

An asset, as I said, is something of measurable value that is owned. So what is measurable value?

The idea of measurable value is that there is a single numeric scale that can be used to compare the value of one thing to another. Something with measurable value can be assigned a number according to that scale, and can be expressed in units. An asset, then, to repeat the definition, is something owned and which has measurable value in this sense. It is something that is owned, and whose value can be compared on the same numerical scale to other things.

This idea of measurable value is very closely related to unit of account which I introduced earlier. The function of providing the scale for measuring value is just the same function as being the unit of account. This function, as we saw, is one of the three definitive functions of money. One of three things such that, if something performs them, that something is counts as money.

The three functions are logically separate from one another. As we saw earlier, something can be a unit of account without performing the other functions of money. Think of the Irish slaves example.

But in a modern society the three functions are normally all implemented by the same thing. The units which express value are normally units of a currency. So a car worth £1000 would be an example of something with measurable value. So would £1000 in cash. Both would have the same measurable value. For this reason what I've called 'measurable value' is sometimes referred to as 'monetary value'. I won't avoid this term altogether, but I do think it is a bit confusing because it seems to assume that whatever is providing the scale for measuring value has to be money – something that fulfills all three functions. But we've seen that isn't the case: think of the Irish slaves again. Irish slaves were a unit of account, they provided a unit for measuring value, but they were not money, because they did not fulfill the other money functions.

In measuring value, just how you assign what number to what asset can be a hard practical problem, and in some cases a hard philosophical one. Perhaps there are some things which have real value but can’t be assigned a number in this way. But these points, interesting as they may be, are irrelevant to the issue at hand. The concept of an asset makes sense if the value of at least some things can be measured in this way. The ownership of assets can be understood in the context of those things. Just how they come to have measurable value, and whether there are things that don’t have such value are interesting questions, but I'm not going to get distracted by them here.

I haven't yet said what fulfils the functions of money. And that's what most of the rest of this piece is intended to describe. For that we need to get back to the concepts of assets and, especially, liabilities.

Liabilities and Obligations: All liabilities imply a virtual IOU.

I’ve said that an asset is something that is owned and has measurable (aka monetary) value. And I’ve said something about what I mean by ownership and measurable value. The next concept to look at is ‘liability’.

To put it very generally, a liability of person X is implied by the right that some other person, Y, has to current or future assets of X. This definition needs some further explanation.

First of all, ‘person’ here could mean real individual people or organisations; both real people and organisations can own assets and have liabilities. (For simplicity’s sake I won’t keep writing ‘people and organisations’, in general I’ll take ‘people’ or ‘person’ to include organisations as well, unless context clearly implies otherwise.)

Secondly, a liability is a kind of obligation, and if you have this kind of obligation to someone, that someone has a corresponding right. Logically, such an obligation and right are two sides of the same relationship. They imply one another. Being in the relationship on one side means you have an obligation, being in it from the other side means you have a right. For example: If a person borrows something, it means that they have promised to pay it, or some equivalent, back at a later time. The creditor has the right to get this at the specified time, and perhaps to receive some interest.

When a right of this kind is fully exercised, and the assets in question transferred, the liability is said to be settled, and it ceases to exist. So a liability comes into existence when a promise to repay is made, and goes out of existence when the promise is fulfilled.

The fact that liabilities can come in and out of existence without the creation or destruction of any real physical objects is kind of obvious, but, as we shall see, is key to understanding the ‘magical’ nature of money, the fact that it too can come in and out of existence without the creation or destruction of any physical objects.

Drilling a little deeper into what a liability is, turns up another important point that will be of value later when we turn to look more closely money.

A liability always involves two assets. There’s the thing that’s owed. That’s one asset that’s involved. But the fact that Y is owed something actually constitutes a separate asset of Y, something that is logically quite different from the thing that is owed. Think of an IOU. The IOU is different from the thing that is owed, and it is an asset of the party to whom the IOU was issued. It’s an asset because it is owned and has a monetary value. For example an IOU might be valued at the same amount as the amount owed, or a bit more if interest is also due.

Not all liabilities involves an IOU in the literal sense of a piece of paper on which ‘IOU’ is written. But a liability always logically implies a corresponding asset of the person owed just in virtue of the fact they are owed something. All liabilities imply a virtual IOU.

Trading and Lending

An economy is based on people trading with one another. Fair trade means swapping assets which have the same value as one another. And to spell this out in detail, what that means is that:

  • At the start of the transaction one party owns something, A, of a given value.
  • The other party owns something, B, of the same value.
  • And at the end the transaction the person who owned A now owns B and the person that owned B now owns A.

It is helpful to diagram these relationships. So let’s start with a playground scenario. Maggie has a marble that Colin wants. Colin has a conker that Maggie would like. So they swap.

In the diagram below I show changes of assets and liabilities. The leftmost column under ‘Colin’ shows his change of assets. The left column under ‘Maggie’ shows her change of assets.

Colin Maggie
+0.5 (marble)
−0.5 (conker)
−0.5 (marble)
+0.5 (conker)

The plusses and minus show changes in their asset ownership. The diagram shows that Colin has gained a marble worth 50p, and lost a conker worth the same. Meanwhile Maggie has lost a marble and gained a conker.

We’re picturing the transaction as simultaneous, so no liabilities arise. But in reality even in a such a simple transaction as this, liabilities are likely to be incurred temporarily. Maggie might give Colin the marble first. But it might take some time for Colin to get his conker. So he promises it to her. At that point, Colin would owe Maggie until his debt is discharged.

During that period, the situation looks like this:

Colin Maggie
+0.5 (marble) +0.5 (debt) +0.5 (debt)
−0.5 (marble)

This diagram shows changes in both assets and liabilities. For each person assets changes are to the left and liability changes to the right, though there are no changes to Maggie's liabilities, so her right-hand column is empty. A plus in an asset colum show an increase in assets, a minus a decrease. Likewise a plus in a liabilities shows an increase in liabilities, a minus a decrease.

What I've labeled "debt" is a liability for Colin and an asset for Maggie. On Maggie's side it is what I called in the previous section a "virtual IOU". On Colin's side it is the obligation to pay what's owed.

The difference between what's going on on Colin's side and what's going on on Maggie's side is fundamentally important to understanding finance. It looks like a simple situation, but the sorts of changes depicted capture almost all the core concepts of finance, money, and banking. So let's spell it out in detail.

At this stage in the proceedings, Maggie has the same number of assets and liabilities that she started with but has exchanged one asset (a marble) for another asset (Colin's debt, a virtual IOU). But Colin has increased both the number of his assets and his liabilities. Maggie’s balance sheet has stayed the same size, but Colin’s has grown. Colin’s has grown because he has acquired a marble and also made a promise and thereby taken on an obligation.

Balance sheets and Income Streams.

The second Colin and Maggie scenario showed Maggie exchanging a marble for Colin’s debt. No money was involved, so let’s now provide scenario where it is. Let’s move from a playground swap to one where a business borrows cash. Suppose Y lends X some cash. We’ll imagine that Y has a £100,000 wad of notes that is transferred to X.

X Y
+100,000 (wad) +100,000 (loan) +100,000 (loan)
−100,000 (wad)

X has gained an asset (wad) worth £100,000. He’s also acquired a liability (loan) with the same value. Meanwhile Y has lost one asset worth £100,000 and gained another, worth the same amount.

Just like the second Colin and Maggie scenario, X’s balance sheet has grown. It’s got a new asset and a new liability. Y’s balance sheet has stayed the same size, but one type of asset has been exchanged for another. A loan for cash.

Why might Y do this? Well, the loan probably has an interest rate, and the cash doesn’t have an interest rate. So the loan produces an income stream. Over time Y’s assets will increase because of this loan, whereas without the loan, his assets would have stayed the same.

Why might X do this? Suppose X makes widgets. The loan may be necessary to buy a widget making machine, so X’s intention is that the business of selling widgets will bring in a stream of income and thereby pay off the loan, the interest, and still leave some profit.

These diagrams show changes in a balance sheet. A balance sheet proper shows all the assets and liabilities of the entity at a given point in time. Here’s X’s balance sheet after X has bought a widget machine. I’m imagining that X’s other assets are worth £50,000.

X
100,000 (wm)
50,000 (Oher assets)
100,000 (loan)
50,000 (Net Worth)

I’ve omitted all the detail, but the diagram is nonetheless helpful in understanding two things.

  • A balance sheet balances.
  • It provides a basis for understanding the kinds of pressures that any kind of business or trading person is under.

The fact that a balance sheet balances is a matter of definition. The net worth of a company is defined as the value the assets minus liabilities. So value of liabilities plus net worth equals the value of assets.

If a company has liabilities whose value is greater than that of its assets, its net worth is less than zero. Equity is another word for net worth – so the company has negative equity. By definition a company is solvent if and only if its net worth is greater than zero. So if its liabilities outstrip assets it is technically insolvent.

This can be serious – if you’re a business you’re not supposed to trade if you’re insolvent. However saying exactly what value a company’s assets have can be a tricky business. And it is possible for a company to be technically insolvent without it having a discernable effect on its ability to trade. For example suppose there’s a dramatic crash in the demand for widgets. Suddenly X’s widget machine is worthless.

But this may turn out not to be a real problem. if X has the means to pay interest on their loans, they could be technically insolvent without anyone actually noticing. If the widget market recovers, and widget machines become valuable again, the technical insolvency would be only temporary.

Avoiding insolvency is not usually as pressing for a business as being able to meet the day to day obligations incurred by its liabilities. If, for example, you are unable to meet interest repayments, or repay loans when they come to term, you’re in immediate trouble.

Any business whatsoever generally has to juggle two things. On the one hand the type of asset which they use to make payments is generally not going to bring in much of an income. On the other hand they need to have enough cash to continue to make payments.

If you make widgets, your widget machines may make up a good proportion of your assets. It is not a great idea to have your assets in cash rather than widget machines, because cash doesn’t bring in much of an income. Still, you need to have enough cash to pay the bills. So running your business means balancing having enough cash on hand to pay the bills with having enough widget machines to make a profit.

Balancing profitability and liquidity is one of the most fundamental challenges for any business. Any business needs to have enough liquid assets – money – to pay its bills, and enough illiquid assets – ones you can’t use to pay bills – to generate profit.

And as we shall now see, this applies just as much to banks as it does to manufacturers of widgets.

Banks

A bank is in the lending business. Its assets are loans not widget machines, though the structure of its business is in many ways similar to that of a manufacturer.

The example of the business loan that we looked earlier had someone lending someone else money in the form of a wad of cash. But that isn’t very common or realistic. And certainly for banks, hard currency is not the main form in which loans are made. If a bank makes a loan, normally this means that the person who gets the loan has their bank account credited with the amount of the loan.

Here’s what happens if X takes out a bank loan of a £100,000 which he’s intending to use to buy a widget machine.

X X's bank
+100,000 (deposit) +100,000 (loan) +100,000 (loan) +100,000 (deposit)

There are critical structural differences between this diagram and one that showed Y lending X some cash. The most important difference is that in the previous example there was a transfer of assets from Y to X. In that example, we can imagine that the wad of cash was physically transferred, but even if the cash stayed in the same place, and only the ownership was transferred, there is, nonetheless, a loss of assets to Y and a gain of assets by X, balanced by the loan which constitutes a corresponding increase in the assets of Y and an increase in the liabilities of X.

In this new case, although the situation is similar for X – he gains an asset, the bank deposit, and a liability, the loan – the situation at the bank is different. There is no loss of assets. Instead the bank gains both a new asset, the loan, and at the same time a new liability, the deposit.

These changes are changes in very specific obligations and rights. The change of balance sheet diagram helps keep track of these. X has the right to withdraw money from his deposit account, the bank has the obligation to provide that money. That’s X’s asset and the bank’s liability. The bank has the right to the interest payments on the loan, and the right to its repayment at term. So the loan is the bank’s asset and X’s liability.

Understanding this difference between the way a bank makes a loan and the way in which a normal person (or company) with cash to lend makes a loan is the key to understanding the creation of money. And it is key to countering many misconceptions about the way that the banking system work. So let me describe the difference again. In the case of a person with a pot of cash that they want to lend, there is an asset – the pot of cash – that they start with before the loan is made, and that is swapped for an for an IOU, virtual or otherwise. They swap cash for debt. Their balance sheet doesn’t change in any other way. In the case of a bank there is no asset that they start with before the loan is made. They create a new asset, the debt, and a new liability, the deposit, both from nothing.

It shouldn’t be too surprising that these things are made from nothing, they are promises. The person borrowing the money makes a promise to the bank. The bank makes a promise to the person. That’s it. On both sides something from nothing. But on both sides too, once the something has been created there are consequences and obligations that it incurs.

Finally, Money

I said earlier that the concept of money which I'm interested in exploring here is that of 'medium of exchange'. Wikipedia has a very nice definition along those lines:

Money is any item or verifiable record that is generally accepted as payment for goods and services and repayment of debts in a particular country or socio-economic context.

There are quite a lot of things that easily and obviously fit this definition, some that fit it but are less obvious, and some which are highly debatable.

The most obvious kind of thing that fits this definition is cash – hard physical currency.

And another kind of thing which definitely does fit it, but which is perhaps less obvious than currency, is a deposit in a bank account. You can use a bank account to pay for goods and services, and for repayment of debts. In fact in any modern economy there’s a lot more of this kind of money than any other kind. 97% of UK money, according to the Bank of England (McLeay et al 2014).

So that’s why the manoeuvre by which a bank makes a loan by simultaneously creating both the loan and a deposit is so important to understanding the way in which money is created in a modern economy. The deposit that the bank creates is money, and, indeed, money in its most common form.

And that deposit is a promise. And like promises in general, it is created from nothing; in a certain context all that a legal person, a human or organisation, has to do to create a promise is make certain statements. Something from nothing. Money is no different.

Keeping Promises Under Control

Because money is type of promise, money, like promises, can be created from nothing. I could stop there – that is why it isn't so very spooky or paradoxical or magical that money is created from nothing. But before I finish, I want to begin to address some concerns that people have about all this.

Perhaps at the back of some people’s worries is the thought that most of us have to strive hard to obtain money, and if you don’t have money you suffer. It therefore seems deeply unfair that banks can magic this precious stuff from nothing.

The most direct response to this concern is that for a bank, money isn’t an asset it’s a liability. A bank doesn’t create money for itself, it creates money in the form of deposits for people and organisations that use the bank. Of course the bank gets something as part of the process of money creation. It gets the asset that it exchanges for the deposit – in the cases we’ve been looking at, it gets the loan. But the asset balances the liability, so, in that sense the bank is no better off after the deposit has been created.

The fact that some institutions are able to create money from nothing does not mean that they can create unlimited money for themselves. You can create a promise from nothing, just by saying the right words, but making too many promises can get you into all kinds of trouble. Money is just the same.

A different concern is that although the money is a liability, the loan is an asset, and that it seems too easy for banks to create such assets for themselves. It is too easy for banks to create loans.

Why should anyone think this a problem? It is worth comparing the situation of a bank to that of a manufacturer. Is it wrong that the owners of the means of production are able to use those means to produce assets from which they profit? If you believe that this is wrong, you owe some explanation of why it is wrong. It is not obviously, on-the-face-of it-wrong: on the face of it, if you own something you ought to be allowed to profit from using it.

Still, in any particular case there can be valid reasons for thinking it wrong that the owner of something use that thing without limit. One such reason would be if its use had bad effects. So, for instance, if an abundance of widgets causes problems for people, then one can argue that there should be fewer made, and that it’s wrong that widget makers are able to produce them willy-nilly. Imagine that a widget is a type of re-recreational drug that causes serious harm to users; or a type of weapon.

The same form of argument applies to banks’ ability to produce loans. Just as a widget manufacturer makes widgets, the profit making assets that banks are able to produce are loans. And just as there’s nothing on-the-face-of-it-wrong with using your widget making capabilities to make widgets, there’s nothing on-the-face-of-it-wrong with using your loan making capabilities to make loans. And so the concern that banks can create deposits and loans could focus on whether their production has consequent bad effects. And this, I think, is where the most serious criticisms of the money making processes I’m describing can be targeted. For example one can legitimately worry that too many loans or too much bank-created money causes various types of economic and social problems.

For example, although money can be created from nothing, most of the things of value that money is used to buy, cannot. So there is a risk that if too much money is created, there’ll be more money chasing the same amount of value, and if the amount of money runs out of control but the amount of things to buy stays the same, you have inflation.

This risk can be overstated. For one thing, the amount of things to buy will respond to the increase in the amount of money available because businesses will be able to make more stuff and offer more services. Only if productive capacity cannot be increased in proportion to the amount of new money, would new money lead directly to inflation. Still it is not realistic to expect productive capacity to increase whatever the rate new money is produced, so inflation is a real potential problem.

That’s not the only potential issue. Another is that huge levels of indebtedness can make an economy fragile because it can become apparent that the promises involved cannot all be kept.

So what I want to do now is describe some of the mechanisms that exist for preventing an uncontrolled proliferation of the liabilities that constitute money. That is not to say that the mechanisms work well enough to provide sufficient control. In fact a good case can be made that many economic problems in our recent and not so recent past were produced precisely because these mechanisms did not provide sufficient control. But to make that case we do have to understand the controlling mechanisms that, in fact, there are.

A Bank Has To Manage Risk

One bad thing that can happen to a business is that it becomes insolvent, the values of its liabilities exceeds that of its assets.

When I described earlier how a widget maker might become temporarily insolvent, I said that the market value of their widget machine could crash. And, in general, a company can go bust if what were previously valuable assets lose their value. When we’re talking about banks, the most important type of asset is a loan. And by the same simple arithmetic, if loans lose their value, a bank can become insolvent.

How can a loan lose value? A loan's value depends fundamentally on its amount plus the interest that it receives. But the value also has to take into account the likelihood of interest and repayment being made. A big loan with a high interest rate isn't worth much if there's little chance of the debtor making their interest payments or making good the loan.

A simple way of seeing how vulnerable a bank is to bad loans, is to measure the ratio of its assets to its equity; its equity, as we saw above, being the difference between its assets and liabilities. To see how this works let's look at a highly simplified bank balance sheet, one stripped down to the absolute essentials of banking, where loans are the only assets and deposits the only liabilities. (Think of the numbers as hundreds or thousands of millions of pounds.)

X
100 (loans) 50 (deposits)
50 (Net Worth)

This bank has a somewhat unusual ratio of assets to equity of 2:1. It's super safe. It's common for a bank to have a much higher ratio, 30:1 or higher in some cases:

X
100 (loans) 174 (deposits)
6 (Net worth)

Why would a bank do this? The reason is that loans produce income, so the more loans you make the more income you have. However there is a big downside to this. In this example of the super safe bank, suppose a chunk of loans go bad – the debtor is unable to pay interest or pay of the loan. Those loans – these erstwhile assets – are then worthless, so the bank's net worth (the difference between assets and liabilities) drops by the value of these loans. Suppose that the value of the bad loans is 7, that's 7% of the banks overall loan portfolio. That woud lead to a drop in the net worth of the bank by a corresponding value of 7, or 14% of its overall net worth.

Consider now the more typical bank. Suppose that it too has a bad loans to a value of 7. But in this case that is 140% of it's equity. It's equity is wiped out, and then some. The bank is insolvent, and is technically bankrupt.

This is one reason that banks can't create loans willy-nilly. The higher their assset:equity ratio the bigger the chance they will go bankdrupt and be unable to trade.

A Bank Needs Sufficient Liquidity to Cover its Deposits

As I’ve stressed, the money that a bank creates is liability to the bank. And a liability is basically a promise. So let’s look at what it means for the bank to fulfill this promise.

In the example above, X has received money in his deposit account. But the reason for his taking this loan is to buy a widget machine. So let’s suppose that he has found a seller who has a widget machine and is prepared to swap it for the sum that X has borrowed.

The diagram below shows this transaction.

X X's bank Seller's bank Seller
−100,000 (deposit_1)
+100,000 (wm)
−100,000 (asset?) −100,000 (deposit_1) +100,000 (asset?) +100,000 (deposit_2) +100,000 (deposit_2)
−100,000 (wm)

After the transaction, X no longer has a deposit, but he does have a widget machine. The seller no longer has a widget machine but does have a deposit.

X’s bank has lost a liability, X’s deposit. And the seller’s bank has increased its liabilities. The seller now has the right to withdraw however much the widget maker cost (£100,000).

That's not all that's going on, as the diagram makes clear. As I said, the seller's bank has a new liability, because the seller is able to withdraw the amount that was paid for the widget maker. But the seller’s bank has no reason to accept this new liability without obtaining some corresponding asset. As a business, accepting liabilities without compensating assets is not a good idea. You’ll go bust if you make a rule of this. So X’s bank has to transfer some asset to the seller’s bank to compensate for the liability that the seller’s bank has taken on.

This point, that the seller’s bank gets a new liability, and therefore needs a corresponding asset is important and sometimes is explained in way that I think is misleading. It’s sometimes said that the corresponding asset is required as a matter of accounting identity, but that can be misinterpreted. From a purely logical point of view it would be possible for the seller’s bank to accept the liability without a corresponding asset. In the sense that it is logically possible, the bank could just accept the obligation to pay the seller. It is logically possible, in the very same sense, for my bank to decide to make me a gift of a million pounds. But the effect would be similar – as a matter of accounting identity such an action would reduce the bank’s net worth, and if it made a habit of doing this kind of thing it would eventually go bust because its net worth would fall below zero.

So, in general, banks don't make gifts like that, and likewise, the seller’s bank needs to acquire an asset to balance the liability that they take on when enabling the seller to withdraw money in her account. For this reason, in the diagram above, I’ve shown that the buyer’s bank has lost an asset worth 100K, and the seller’s bank has gained one.

There are several types of asset that can play this role. And they are important. In fact, it’s little exaggeration to say that the history of modern banking is a history of the kinds of asset that play this role. Assets that play the role of compensating for transfers between banks are important because they provide liquidity to banks. They are what enable banks to keep operating, day to day, just as cash-flow keeps a widget making business going. If a widget manufacturer can no longer meet its day to day obligations, regardless of its overall solvency, it’ll go out of business. The same is true of a bank.

At this point I don’t want to get into too much detail about the nature of the assets that can play this role, but for concreteness I’m going to describe one important type: deposits at a central bank.

The central bank is a bank for banks, and only banks have accounts with it.

Here’s how a central bank is involved in moving what I’ve labelled "asset?" from bank to bank when money is transferred from an account in one bank to an account in the other.

X X's bank Seller's bank Seller
−100,000 (deposit_1)
+100,000 (wm)
−100,000 (deposit_at_cb_1) −100,000 (deposit_1) +100,000 (deposit_at_cb_2) +100,000 (deposit_2) +100,000 (deposit_2)
−100,000 (wm)

Central Bank
−100,000 (deposit_at_cb_1)
+100,000 (deposit_at_cb_2)

This is the same diagram as the previous one, except I’ve spelt out the nature of the asset transfer that balances the transfer of deposits. After this transaction has completed and cleared, the value of X’s bank’s deposit at the central bank has gone down. The value of the seller’s bank’s deposit at the central bank has gone up. The Central bank meanwhile loses a liability to X’s bank and gains an equal liability to the seller’s bank, so overall the value of its liabilities doesn’t change.

As I mentioned, an account at the central bank is but one example of an asset that can be easily transferred from one bank to another. There are others. By contrast many types of asset cannot be so easily transferred. For example loans, although they are assets, cannot be moved about in this way. In the terminology of finance, loans are not ‘liquid’ but central bank deposits, and other kinds of asset, are.

A bank should have enough liquid assets to cover expected withdrawals from deposit accounts. If you’re a bank and the proportion of your assets that are liquid is too small in relation to the number of deposits that people have, you can easily hit trouble. Depositors will want to withdraw their money, but you won’t have the liquid assets needed to support those requests.

So this is a second constraint on the creation of loans. Loans are not liquid, but the creation of loan implies the creation of a deposit, and that deposit is likely to be withdrawn, and so in order to support the creation of that deposit the bank needs corresponding liquidity. Banks can’t create loans willy-nilly because they’ll run out of sufficient liquidity to cover withdrawals.

What Next?

This is much longer than I'd intended it to be, but I hope that, as promised, I've delivered as account of why we can say that money is created from monthing, and, in addition I've started to explain why this isn't as weird or dangerous as it might first appear. Let me summarise:

  1. Money is a bank liability, is therefore a promise, and so can be created from nothing.
  2. You'll end up in mess if you make promises you can't keep, likewise there are factors which constrain the amount of money which banks can create.

Whether there are sufficient constraints on money creation is quite another question, and there are good reasons for thinking that the answer is 'no', see for example Admati and Helliweg (2014).

It might be interesting to look at some of the alternatives that are proposed to using bank liabilities as money. Perhaps the most famous is that we should return to using gold as money. I strongly suspect that when such proposals are analysed correctly their downsides and risks far outweigh any potential advantages. But I'd like to see this spelled out in more detail.

But here's what I think is the most interesting place to go next. Private banks are not the only institution which creates money. When the government spends more than it takes in tax, that deficit too is money creation. When this happens the central bank is obliged to sell bonds of the same value as the money created. There is a lot I don't understand in this process. Why the bond sales, for example? Another question: Why do governments feel so constrained by deficits? And why the obsession with the summed values of bonds that people have chosen to buy, aka government debt?

References

  • Admati, A and Hellwig, M (2014) he Bankers' New Clothes: What's Wrong with Banking and What to Do about It. Princeton University Press.
  • Graeber, D (2014) Debt: The First 5,000 Years. Melville House Publishing; 2nd Revised edition edition.
  • McLeay, M, Radia, A and Thomas, R (2014) Money creation in the modern economy. http://www.bankofengland.co.uk/publications/Documents/quarterlybulletin/2014/qb14q1prereleasemoneycreation.pdf
  • Roth, S (2016) Isn't it time to stop calling it "The National Debt". http://evonomics.com/isnt-time-stop-calling-national-debt/

Published on Apr 17, 2017