A bank approves a loan: where does the money come from?

In the naïve picture of the banking system, banks strike a balance between savings invested in them by savers and the loans requested of them by borrowers. The money loaned to borrowers is the same money invested as savings, and some of the interest paid by the borrowers is passed onto the savers. In this model the bank is effectively a middleman.

In the real economy this isn't what happens. Because money is no longer tied to physical currency, there is no longer necessarily a direct link between money saved and money loaned.

And if the money for the loan, which once granted is indistinguishable from any other money, just comes from a corresponding negative entry in a bank's balance sheet: where does it actually come from? Is the money being lent out 'real'? Is the act of a bank granting a loan intrinsically different from, say, you or I doing so?

• watch this: youtube.com/watch?v=KvpbQlQwl0A Jul 16 at 0:28
• @Timkinsella Thanks for this - it and the other videos in the series are the closest I’ve got so far to an answer. It’s still not clear to me what happens with loans defaulting, though: if less money is destroyed (through repayment of loans) than created (through approval of loans), doesn’t that imply the rate of increase of money supply is related to the rate of default?
– JCW
Jul 16 at 11:37
• You might like this too youtube.com/watch?v=CI5CFQXJxcA
– Mick
Jul 16 at 11:43
• @JCW Yes. When a loan is made, money is created. When a loan is repayed, money is destroyed. Every default means the money that was initially created by the loan will remain in the economy indefinitely. Jul 16 at 19:53
• @Timkinsella Has anyone explored the macroeconomic implications of this? It seems like a big deal, and I’ve never seen it mentioned before, let alone explored. It certainly isn’t common knowledge. It presumably implies that in the long run, money created by defaults will come to dominate? Does this imply that granting riskier loans and accepting a higher incidence of defaults generates economic growth?
– JCW
Jul 16 at 21:08

Banks create real money simply by approving a loan.

The intermediary "theory" is a misconception, at least regarding money creation as it works today as you point out yourself. Banks don’t lend out customers deposits. On the contrary: By approving a loan, a bank, say Bank A, creates a deposit of the same amount in the customer’s account. For all intents and purposes this deposit is already real money. In fact, this so-called broad money is used for the vast majority of transactions in a modern economy (direct debits, bank transfers, credit card purchases, etc.).

As a result of this money creation, the balance sheet of bank A grows. There is a claim on repayment of the loan on the asset side, and the liability side reflects the new deposit which A owes to the customer. But A’s reserves - the kind of money only the central bank creates, and which is used by commercial banks to settle accounts between them, remain unaffected by that process. Only when the customer transfers some of the money to another bank B do reserves come into play. A will transfer its liability (the deposit) to B and, to make up for it, it will also transfer reserves of the same amount (an asset). Although A has now less reserves, and B more, the total amount of reserves in the banking system has not changed, other things equal. So, banks can create real money (the money used in commercial transactions) simply by that process, and reserves don’t play a role in the process itself.

Effects of making a loan on balance sheets (without reserve requirements) Source: BoE quarterly bulletin extract (2014)

However, if in place, reserve requirements (when banks need to hold a certain ratio of their deposits in reserves or cash), do limit the amount of money that can be created, but are not involved in the process of how money is created by commercial bank. Reserve requirements, even if they exist, have become much less relevant over time, partly because of securitization (of loans/deposits, think about lowest-quality mortgage securitization ahead of the financial crisis), which allows them to restructure their balance sheet in favor of more reserves.

That doesn’t mean the central bank can’t steer the creation of money or do it itself. On the contrary, monetary policy is all about balancing the relationship between inflation, money supply, and interest rates (the price of money). And of course, the central bank can affect the amount of money more directly via quantitative easing, but that was not the question.

Capital requirements are designed to curb systemic risk and, possibly, to limit the speed of money creation. If a bank accumulates profits into capital, it can still grow its deposits and hence create ever more money, the speed of which is only limited by the speed at which it can grow its capital.

What will limit the amount of money that is created are the banks’ profit considerations in the context of competition and macro-prudential regulation. You could argue that banks can, and do, create money out of “thin air” when they approve unsecured loans (credit card debt, overdrafts), because there is no collateral backing it up. But in many cases, especially for larger loan amounts, banks will require some high-quality collateral to approve a loan (for example the rights to the home for a mortgage loan).

As banks increase their loan portfolio, remaining loan opportunities will eventually become too risky relative to interest rates banks can charge in a competitive environment (or, in other words, borrowers as a whole run out of high-quality collateral). At that point banks will stop lending and, hence, stop creating money. Note that this risk is perceived risk, and money creation by commercial banks will ultimately depend on how they perceive risk. Macro-prudential regulation, such as caps on loans-to-value ratios impose limits to the risks banks can take from the outside, and hence limits money creation in that way.

If an individual person were granting a loan it would be different for a variety of reasons. Most importantly, unlike banks, you don't have access to reserve money and cannot borrow reserves from other banks or the central bank when people actually want to see their money right away, which will always be the case in that situation. So, you wouldn't be able to create money.

References (all focusing on the UK, but most mechanisms are general):

• In addition the first bold sentence is wrong according to your own answer, even though the rest of the explanation is sound, since as you explain in your answer banks cannot create money out of thin air, they can only do that within the constraints imposed by central bank from capital/reserves
– 1muflon1
Jul 16 at 18:36
• -1 by me the first sentence is blatantly false, I don't care how well rest is written. You should read Rendahl & Freund voxeu.org/article/banks-do-not-create-money-out-thin-air Jul 16 at 18:40
• @csilvia: Just because someone else says the opposite doesn't make it false. What's wrong with my reasoning? Btw, I admit that "out of thin air" is not scientific, open to interpretation, and may have a provocative touch.
– BrsG
Jul 16 at 18:48
• @BrsG no how is affected by this since it suggests that private bank can only produce money from high powered money. Saying its created out of nothing is incorrect, if it is created using some other instrument. Take an analogy, if firm producing totems can manufacture 10 totems out of piece of wood, the totems are not created out of nothing just because in one time period firm only uses 50\% of its wood inventory and creates only 5 totems and then uses rest of that inventory in next accounting period. The same holds for capital/reserves. Banks can only create money from high powered money
– 1muflon1
Jul 16 at 19:23
• @csilvia: edited my post to explicitly encompass voxeu.org/article/banks-do-not-create-money-out-thin-air.
– BrsG
Jul 19 at 9:23

where does it actually come from?

The money comes from the reserves (note not in 1:1 ratio necessarily - the ratio of reserves to money is typically very small actually) provided by central bank (or possibly from other private bank that happens to have excess reserves/capital), and the reserves themselves are created out of nothing by government decree by central bank (hence why it is called fiat money). If a bank lends out money, and the lending is not made from deposit, then it has to ask central bank for enough reserves to maintain its reserve requirements (or in present day capital requirements since reserve requirements were abolished but capital requirements work in analogous way).

You can have look at McLeay et al 2014 for more detail (although the paper is bit out of date nowadays due to abolishment of capital requirements and in some places uses bit non-standard terminology).

Is the money being lent out 'real'?

Of course it is. You can use the money for everything you can use coins or notes in circulation for. In addition, ultimately all fiat money comes from government whether it be central bank (or in some countries like US, both central bank and federal government since mint and bureau of printing is technically part of US treasury).

Is the act of a bank granting a loan intrinsically different from, say, you or I doing so?

In what way? Banks can endogenously increase money supply by asking central bank for additional reserves, if they face high demand. An individual cannot do that. Obviously, there is also more complexity in bank loan than in individual loan because of all the institutional context.

• @JCW most money is digital not in cash, if banks gives out loan it does not actually hand you bag with a \$sign on it full of bills, it sends it to your account. Also, supply of physical cash is not really limited, banks can order more of that stuff it there is insufficient of it to circulate (here is an explainer on how the cash works newyorkfed.org/aboutthefed/fedpoint/fed01.html). If more cash is needed more can be always printed out, but nowadays people typically just keep money electronically most loans are never even cashed in the literal sense of that word. – 1muflon1 Jul 15 at 19:23 • @Mick no what you say is completely incorrect if you would actually read literature, if we want to go into detail correct way would be stating \$1000 come from \$10 (or whatever reserves are required) not from nothing since the 97% (or whatever the true share is) cannot exists on its own without the 3% and is derived from it. So yes money literally comes from reserves not in 1:1 ratio, but I never claimed it did – 1muflon1 Jul 16 at 10:34 • @Mick for example no matter what’s the borrowers future streams of cash flows are if banks has no excess reserve, and central bank is not playing ball and refuses to create new reserves, no new money can be created, loan cannot be given out otherwise bank would violate capital requirements regulations. Only if the bank has sufficient capital/reserves new loan can be extended. Of course not necessarily in 1:1 ratio, but reserves are necessarily – 1muflon1 Jul 16 at 10:47 • @Mick sure for loan to exist there must be demand for loan, but the money for that loan come from high powered money not from nothing that’s the point – 1muflon1 Jul 16 at 10:54 • "generally banks don’t have reserves just laying around" may have been true before 2008 but the position may have changed: see fred.stlouisfed.org/series/TOTRESNS Jul 16 at 11:28 The money is """real""" in the sense that it's backed by "real money" even though it's not "real money". Think about this: You go into a bank and ask for a \$1,000 loan. They say yes. They update their database - now their loans are \$1,000 more and your balance is \$1,000 more. So far, no "real money" has been involved.

But then, you go to the teller and you say "I'd like to withdraw \$1,000 please." Now they have to give you$1,000 of cash. And they aren't allowed to print that cash - that cash is "real money" to them.

Maybe you don't want cash. Maybe you want to buy a cheap car, so you use your card, and now your balance is \$1,000 less and the car seller's balance is \$1,000 more. And now the car seller may go to the teller and ask for \$1,000 of real money. The fact that your bank balance may be converted to real money at any time is what gives it a real money value. In any case, after the withdrawal happens, the bank is now down \$1,000 of real money, even though they have the same total balances in accounts, and \\$1,000 more of loans. The bank has to keep a close watch on the account balances people have, versus the amount of real money they have. If their account balances are too high compared to their real money, they can't make more loans.

If they do make more loans anyway, it puts them at risk of bankruptcy, and they don't want to do that because bankruptcy makes their shareholders lose money.

• Thanks for your reply. I understand this presentation of the concepts. But I still don’t quite see how it can work like this in practice. We know that banks don’t actually have enough assets to convert their entire balance sheet simultaneously to ‘real’ cash, so your ‘The fact that…’ line seems wrong? And we also know that some non-negligible amount of debt is never repaid. That is, the bank would on average release more ‘real money’ into the economy than it gets back - except that it also collects interest payments. Does the bank therefore rely on total interest exceeding defaults?
– JCW
Jul 15 at 17:22
• @JCW Well, yes? Of course they rely on interest exceeding defaults. Otherwise they'd gain nothing from the loans. Jul 16 at 9:26
• @JCW And yes, there are strict regulations on banks to make it so they're extremely unlikely to run out of real money, and there are other regulations to make it so in the unlikely even they do run out of real money, the government can print some for them (but they have to pay it back later!). Before these regulations were put in place, a bank collapsed and all its customers lost their savings every week. Jul 16 at 9:27
• Of course they expect interest to exceed defaults at the point of granting the loan, but given what we know of the economic cycle, it is almost certain that there will be periods in the lifetime of a typical mortgage (say) when aggregate defaults will exceed aggregate interest. Is this why banks reduce lending during financial downturns?
– JCW
Jul 16 at 11:11
• @JCW Sure, and also why they build up savings when things are good, same as any business that doesn't want to go bankrupt. I think that part can just be analyzed as business profits/losses. Jul 16 at 11:15

Each bank has a Chart of Accounts including accounts that are designated Assets, accounts that are designated Liabilities, and the difference between Assets and Liabilities is Owner's Equity or just Equity.

This four page pdf paper shows a simplified balance sheet and income statement for a fictional bank:

https://www.richmondfed.org/~/media/richmondfedorg/publications/research/economic_brief/2012/pdf/eb_12-03.pdf

where Assets equal Liabilities plus Equity in the balance sheet. The reserves for loan loss account (also known as the loan loss reserve or LLR) is considered to be a contra-asset account. If one makes this account equal to zero then the assets would increase by that amount and the Owner's equity would increase by the amount in the LLR. So the loan loss reserve is a way to anticipate an expense charge against equity claims in the event that there is a default loss in the loan portfolio. Therefore equity claims on banks are reduced when loan defaults occur. Deposit levels are not directly coupled to loans or defaulted loans due to the equity market which drains deposits to create equity claims on the Liabilities and Equity side of the bank and aggregate Bank sector balance sheet.

This ten page pdf paper describes how banks create money and then also borrow money to keep interbank payments flowing in the reserve payment system: