Like many noneconomists who read about economic history and finance, I always remain somewhat baffled by the opaque generation of currency under the federal reserve system. My understanding is that the banks "loan" money to the government, which prints a federal reserve note or IOU signed by Treasury. But then what?

How do the banks actually receive federal currency and/or credit... and determine the amounts they receive? My understanding is that banks can effectively receive as much money as they can "justify" by credit contracts, ratings, and a given ratio of leverage or fractional reserve. Apart from market forces and assumed prudence, are these the only factors that legally limit the amounts of money banks can receive? Can banks essentially get as much money as they say they can securely loan?


3 Answers 3


The phrasing of your question (the same is true for your description of your understanding) suggests to me that what you need to understand about money creation in banking is much more basic than "steps and contracts". As a full treatment of money creation would not be appropriate on a Q&A site (it'd be really long!), I will instead recommend that you first read about the different types of money (specifically, the difference between base money and broad money), and then that you read "Money Creation in the Modern Economy", which will answer your question about how banks create money.

The short answer to your question is that banks don't "receive" money to loan out as a part of any formal arrangement, rather, they are allowed to lend out a large fraction of the money that is deposited with them. Importantly, when they give someone a loan, the money that is loaned will tend to be deposited at a bank (because banking is a mostly closed system), which gives the bank that receives the deposit the option to loan out more money. That's how money creation happens— loans that, through the magic of double-entry accounting, result in the bank swapping one kind of asset (reserves) for another (an outstanding loan), while the borrower suddenly has a new liability (the loan they owe to the bank), and a new asset (cash).

When we talk about money creation, we are not saying that banks literally create money in the sense of asking that some be printed up for them, it's that the banking system acts in a way that generates total liabilities (or claims on others) that are much greater than the monetary base. Because all money is really just claims on others, this is, in effect, money creation.

  • $\begingroup$ Thanks. I'll check those sources. Yes, I'm sure "money creation" is not the right way to put it. I sort of grasp the fractional reserve and double-entry concepts. I guess I'm trying to figure out the role of banks in determining the overall money supply. At some basic level, is the "judgment of banks" as to creditworthiness the essential determinant of money supply? Perhaps I'll try to reframe the question when I know more about the tiers of money supply and figure out what I actually want to ask. $\endgroup$ Nov 27, 2015 at 17:59
  • $\begingroup$ @NelsonAlexander You're welcome! Yes, the judgment of banks does have a major role in determining the total money supply (it's possible that if banks do not see profitable lending opportunities once accounting for risks, they will not lend), though at other times they're constrained by other factors, like reserve ratios, leverage ratios, or liquidity coverage ratios. $\endgroup$ Nov 27, 2015 at 19:31
  • $\begingroup$ I think of it a different way (but I'm not an economist): Banks can't create cash, but they can create bank accounts. The only reason that banks get to create money is the fact that everyone treats bank deposits as money. If people only accepted cash as money, then banks could not create money (any more than the stock market can). $\endgroup$
    – user253751
    Jul 29, 2019 at 3:46

There are two types of money:

  • Government money or the monetary base (MB) which consists of paper dollars, electronic dollars, US notes and coins.

  • "checking account money" (aka demand deposits). We have more in checking (in the aggregate) than MB and we accept checks as money so they are money. The most narrow definition of the is M1, but a more accurate/broader definition is M2 or even M3.

The process by which both enters the economy is very different!

Let's start with how MB is created. First it is very useful to study the Fed balance sheet to get a conceptual understanding how this works. For the Fed, when they create money it is a liability. To the individual that holds the money it is an asset. Let's monitor a transaction that introduce 100k into the economy. The Fed sees the rate at which banks charge each other for short term loans change from 4% to 4.5%. Say the target Fed Funds Rate is 4%. The Fed tries to restore 4% by adding money to the banking system. So they go to the "Open Market" and find a bank (only a special one called a "Primary Dealer") that will sell them an asset for 100k (treasury REPOs are usually the asset of choice). The Fed creates a 100k liability (Fed deposit) and uses it to acquire the security (now an asset to the Fed). The Primary dealer now owns 100k in electronic dollars (or a 100k deposit at the Fed that doesn't exist before). The 100k represents a 100k increase in the MB. The Primary dealer of course will spend or relend this to the rest of the economy and that is how MB is created.

So how is bank money created? When you deposit say 22k in cash at a bank, they keep 22k in reserves (cash or electronic dollars), and they give you a 22k claim on those reserves. But money hasn't been created yet. It is when a bank issues more claims on reserves than it has, that bank money is created. So if a banker has 232k in reserve assets and 232k in deposit liabilities, but notices that on average all reserves are never completely redeemed. So say they overbook their reserves by selling some to another bank. Say 32k. The bank now has 32k in say treasury assets, 200k in reserves, 232k in deposit liabilities. Money has been created. This is inherently unstable because the bank is issuing more promises than can be kept and central banks work very hard to subsidize this system so there is never a "bank run". Whether they should is another question (it can cause inflation and instability).

Then some clarifications in regards to your specific points.

Banks do not mostly loan directly to the government, but they do so indirectly. A lot of bank money is used to purchase securities from the secondary market, but not so much the primary market. Certainly bank money is used to purchase many other things like home loans.

Federal reserve notes and coins (M0) are not significant to determining monetary policy because they ommit electronic dollars (or deposits at the Fed). The Fed will gladly switch electronic dollars to paper or vice versa, so the difference between the two is not significant. M0 (physical money) + federal reserve deposits (electronic money) = MB all government money.

Practically speaking the banks do not dictate to the Fed how much MB they receive (which in turn determines how much bank money they can pyramid onto this new MB). A modern explanation of how banks works figures in the influence they cause to the "open market" which means to an extent they can determine how much the Fed gives them. This is very complex and unnecessarily obscures sound underlying concepts of how fractional banking works. I wish many economic students are not told about this first as it really confusing things. I'll try to explain.

The Fed's favorite weapon of choice for introducing MB into a economy is through the "open market". They attempt to influence the Fed Funds rate to be an arbitrary interest rate. If they banks charge each other more for short term loans, this rate goes up and they will induce the Fed to create more MB. In a modern system, banks heavily borrow and lend from each other. Your common bank then in turns borrows and lends from super banks (primary dealers), and super banks (primary dealers) borrow and lend to the Fed. In a "modern system" the banks are not concerned about reserves...because they know indirectly they will be provided (the basic concepts of fractional banking are still valid though).

So say a bank sees a great opportunity to acquire a 100k interest bearing asset. They just buy it (unless their credit rating is suspect). And then borrow the reserves they need to buy it from another bank. If reserves are scare, the rate at which reserves are charged goes up, which induces the Fed to add more reserves to the system. This is complex and again obscures very sound technical concepts of how banking works.

If the banks got together as a cartel and fixed the Fed Funds Rate, then yes, there is not limit to the extent to which the open market could be manipulated to provide reserves to bank (well more so than currently). A logical concern is that the Fed is already buying overpriced assets and getting churned by excess/manipulated buy/sell volume which means your average taxpayer is hurt.

  • $\begingroup$ Thanks you for this very good, labor-intensive answer. The Fed operations in paragraph 6 are illuminating and usually missing in explanations centering on bank money. However, it is unavoidably confusing. Economics seems all trees no forest. And, as with areas of physics, the concepts deteriorate if you don't use them regularly. My only other question might be recommendation for a book with a very "reductive" or "all forest" perspective on state-bank-credit operations and generation of national currencies. $\endgroup$ Nov 29, 2015 at 1:56
  • $\begingroup$ Part of the problem is that a lot of "experts" don't actually know that much about how central and private banking works. I have not found that perfect book...perhaps the best I can recommend is: wfhummel.net Meant for beginners, but the author probably knows more about banking/money than most bankers and Fed officials. He is a little dogmatic though and uncritical of the status quo (like fractional banking and the Fed). Good info though. $\endgroup$ Nov 29, 2015 at 14:11
  • $\begingroup$ A good conceptual understanding of how fractional banking works, can be obtained from reading books from the Austrian School of Economics. Murray Rothbard is a great example and many of his books are free online. Here is a fine one called: "The Mystery of Banking". islamic-finance.com/library/Rothbard_Mystery_Banking.pdf My one complaint is that while he does a great job of analysing how banking works, he uncritically advocates the gold standard which I disagree with. $\endgroup$ Nov 29, 2015 at 14:15
  • $\begingroup$ I'll note that (by my understanding) the bank can keep all its promises, as long as it doesn't have to fulfil them all at once - if it had to, it could gradually wait for its loans to be converted into cash, then give the cash out, without issuing new loans. The problem is if everyone wants their cash all at once and doesn't want to wait for those loans to be repaid, and doesn't want to accept payment in loans instead of cash. $\endgroup$
    – user253751
    Jul 29, 2019 at 3:48

A lot of the confusion around this subject is created because of a problem of definition within economics, combined with a long term process that has changed the actual form of money being used in the economy over the last several centuries. That and some extremely badly written textbook descriptions that can ultimately be blamed on an a misquoted simplification provided by Keynes to the British Parliament in 1931(Macmillan Report).

Banks generally don't loan money to the Government. They can, the various parts of the Government can have bank accounts just like anybody else, but usually the closest they get is buying treasuries as an asset. In today's monetary economies - this changes with time and place - physical money is printed on demand, and while it's not quite an error term, it is a very small fraction of the total amount of actual money being used.

Ever since the invention of cheques specifically, and in general the ability to make direct transfers between two deposit accounts, it is the money represented in a deposit account that has come to be the actual unit of monetary transfer used in modern economies.

Money in deposit accounts gets created in two different ways - one is a deposit of actual cash money, and the other is any time a loan is made. (It should be noted here that deposit account money is also destroyed when a loan is repaid, or as part of loan write-off, so the actual growth (or contraction) in the money supply depends on the rate of new lending versus loan defaults.)

For any non-economists with an engineering background - the actual relationship between physical cash and bank deposits managed by banks is better known as 'statistical multiplexing'. Back in the day (around the 15th century), faced with the alternative of going to a known location to withdraw a heavy metal to use for monetary transfers, in cities that would remain innocent of organised police forces for a couple of hundred more years, goldsmith customers chose to transfer their deposit chits between each other, rather than the actual gold they represented. The early gold smith bankers on their side observed that they could get away with some quite significant amount of short term lending against the gold, since most of the time the gold wasn't being used. Quite rapidly - by the end of the 19th century the American Economist Dunbar is reporting that over 90% of all transactions are being made through the banking system - the new form of money supplanted the old.

As a system it has quite a few advantages - and next to the post office is one of the earliest forms of wide area network ever invented. It has some significant disadvantages too - the biggest being how to regulate it to not expand or contract too fast. Regulatory frameworks have changed a lot over time, and place, and are still a work in progress. For the USA at the moment, reserve ratios play little or no role (the reserve requirement only applies to Net Transaction Accounts (chequeing accounts essentially). There are separate controls introduced by the Basel Treaties that regulate lending as a multiple of bank capital, which are probably the most significant limit on bank lending in the US at the moment, alongside custom and practice limits on the amount of loans that people are allowed to take out.

  • $\begingroup$ Could you provide a reference on this idea that bank deposits and physical cash is somehow related to statistical multiplexing? Strangely enough, I'm an economist who started out in computer engineering, and I don't see any similarities between the two, nor have I ever seen such a link suggested by someone other than you. I'm open to the idea that there are some sort of structural similarities, but I'd really like to read more. $\endgroup$ Nov 27, 2015 at 23:58
  • $\begingroup$ Unfortunately I can't - I keep wondering if I should write this up btw. I have presented at a couple of conferences where I've mentioned this, and had several suitably distinguished (well, they were old and had beards :) professors agree with me - fwtw. If we think of physical money as the thing that is statistically multiplexed, and note that it's actual purpose in the banking system is to be transferred between banks as a semaphore for a deposit transfer, then I think the relationship is fairly clear - or perhaps most importantly that's why it worked. Consequent instabilities included. $\endgroup$
    – Lumi
    Nov 28, 2015 at 0:18
  • 1
    $\begingroup$ Yeah, that's probably worth writing up. I'd read it :) $\endgroup$ Nov 29, 2015 at 21:22
  • $\begingroup$ @dismalscience Isn't it obviously the same basic concept? You can promise people more than you have, on the assumption that they won't all try to use it at once. (For historical reasons, this is not considered a felony, even when you do it with money) $\endgroup$
    – user253751
    Jul 29, 2019 at 3:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.