Fractional Reserve Banking: a myth?

Great Debates
1

I’m going to argue that the conventional story* of how fractional reserve banking works is wrong. Specifically, that it’s inaccurate and unhelpful.

The part of it that is right is that it acknowledges that banks create money by lending.

But there are a number of problems with it. One is that it fails to distinguish between commercial bank money, and government money. The difference is critical, and the failure to emphasize it is misleading.

Secondly, it says it’s the reserve requirement which restricts lending, when the reserve requirement actually applies to deposits, not to loans. The set of regulatory rules that apply to lending is called the capital requirement.

Third, it states (or at the very least implies) that the money banks lend is the same money customers deposit. The truth is, when banks lend, they simply create new money. (Commercial bank money, which is the only kind of money they’re capable of creating.)

Fourth, it gets the order wrong. In the standard story, banks obtain deposits so that they can lend. Banks lend when they have credit-worthy customers (and meet capital requirements). If the amount they have in their reserve account is not enough to meet the reserve requirement, they obtain the necessary reserves when they’re needed. (Remember: the reserve requirement applies to deposits, not to loans.)

A hypothetical example: a loan officer at a bank gets an application for a mortgage. She does what loan officers do: she reviews things like appraisals, credit scores, inspections and bank records. She makes a decision about whether to make the loan, and at what interest rate.

If the loan is approved, the bank cuts a check, and the check is deposited at some bank, somewhere. Since the check is going to the seller (of the house) it will be deposited at whatever bank the seller uses.

The check will take some time before the funds become “available” - in other words, for the transaction to settle. During that time, the Fed will get instructions to transfer funds from the Federal Reserve account of the first bank, to the second one. (If both the buyer and seller use the same bank, amount in the reserve account will remain the same.) When the check settles, the amount will show up as a deposit in the seller’s bank account. The amount is new money, that didn’t exist before. It is the loan that created the deposit. There would be no new deposit, unless a bank, somewhere, first made a loan. If, at the end of the day, a bank somewhere has acquired additional deposits (and a bank, somewhere, will have acquired additional deposits, if there’s a net increase in commercial bank lending) it must get additional money to meet the reserve requirement, if it doesn’t already have it. The quickest, easiest way to do it is in the federal funds market. The federal funds market is where banks go to borrow money (federal funds) from other banks.

This is where the distinction between government money and commercial bank money becomes important. Commercial bank money exists in only one form: deposits at commercial banks. And it exists only as numbers kept in banks’ computer systems. All other kinds of money - currency and Federal Reserve deposits - are created by government. (Which is not to say that government money is owned by the government: it’s owned by banks and people. Reserve accounts and currency are liabilities on the government, in the same way that commercial bank accounts are liabilities on commercial banks. Government money is owned by people and banks, in the same way that commercial bank deposits are owned by depositors.)

There are other differences. One is that commercial bank money is backed by government money - you can, if you want, go to your commercial bank and close your account and get currency. Government money, on the other hand, is backed by nothing. You cannot go to the Federal Reserve (assuming you had an account at the Fed, which of course you don’t) and exchange your dollar for anything but a dollar. Which means that while a commercial bank - or even all commercial banks - can go broke, from lack of currency, for example, when customers are demanding it - the government can’t go broke. The government’s ability to create money (unlike commercial banks) is unlimited.

If the Fed, for example, decided to reduce the money in Federal Reserve accounts to zero - which it would do by selling off its assets - it would bankrupt all commercial banks. Without reserve accounts (or more specifically, without money in reserve accounts) banks would not be able to conduct transactions, or meet regulatory requirements.

Another difference is that commercial banks are subject to financial panics and bank runs - and the more they lend, the higher the risk becomes. Central banks are never subject to financial panics or bank runs.

Since banks, collectively, cannot create reserves - currency or central bank money - the government (specifically, the Fed) must supply them. If bank 1, for example, acquires reserves by luring customers from bank 2, the total amount of reserves has not increased. Some reserves have simply been transferred from one bank to another.

If, at the end of the day, the total amount of reserves required by banks is more than the amount available, the Fed must choose between shutting banks down - for failing to meet a reserve requirement the Fed itself has imposed - or providing the reserves the Fed itself requires them to have, and which only the Fed can provide. Since the Fed doesn’t shut down otherwise healthy banks under these circumstances - and it would be stupid public policy if they did - as a matter of course, the Fed will provide, and must provide - whatever reserves banks require.

The reserve requirement, therefore, has no actual impact on bank lending or commercial money creation at all, much less in the way the standard story says it does.
*The conventional story:

2

This is an interesting thread, so I’m just putting down a marker to remember it in future.

As an aside (maybe helpful, maybe not) I believe early banks would take deposits in gold, and lend it out in exchange for IOUs (which were backed by other assets, such as land, by the borrower). After a while the demand was bigger than the gold in the bank’s vault, and so the bankers started giving ‘Promises to Pay’ the amount of gold, which were exchangeable. Hence the total amount of ‘Promises to Pay’ and ‘IOUs’ of gold were bigger than the actual gold in the vaults.

If that is true, it may have some bearings on the debate. If not, then please ignore.

3

For people having a hard time understanding the OP, here is one of my posts in another thread that lays out the issue here with different wording.

Short version: although there are various small mistakes, the basic idea in the OP is correct. Edit: I would say also the title of the thread is ill-chosen.

4

Was going to post something, then realized I didn’t know what I was talking about quite enough for the thread.

5

You probably did. Many people seem to consider “fractional reserve banking” to be some mystery or fraud akin to David Copperfield making the Statue of Liberty disappear. In fact, paper money is much simpler than that.

One night I overdid it at the local tavern, and the stripper who was lap-dancing me wouldn’t even give me my billfold back. I scribbled an IOU for the bar tab out to Joe the Bartender. bump says “Hey, that Septimoo is an asshole but he might good for the Benjamin.” Presto! Paper money has just been created (although the $100 note might sell at a discount, just as U.S. greenbacks did in the mid 1800’s).

A check drawn on Citibank is little different in principle from my scribbled IOU, one difference being that banks and the FRB must have real assets backing up their liabilities. The check is much more easily negotiated than my scribbled IOU, and the bank has submitted to Federal scrutiny. (Non-chartered banks can have zero reserve in effect, rather than fractional!)

I hope one of the Board’s economists can give us a strong objective look at MMT. Worrying about [del]David Copperfield[/del] fractional-reserve banking is beside the point.

6

Basically true.

But banks aren’t necessary for this. Even without the assistance of a financial institution, people could write a sort of proto-check. “I will pay you 100 GP in three months time” and then sign it over. The person who received this proto-check might then be in need of immediate purchasing power without the hassle of waiting three months for the golden coins. But they have the paper, don’t they? That paper is a promise to pay, and therefore valuable. They can then sign this paper on to another designated person (endorse it and use it for payment) rather than keeping the bill to maturity. Then the next person could endorse it on again. And again.

Some of these bills of exchange accumulated more than a hundred endorsements before maturity.

And at the end of the time period? When the bill came due for maturity? Hell, gold might still not be necessary. The mature bill might be replaced with yet another new bill, promising payment in another three months, what we call today “rolling over” the debt. Rather than the dense precious metal circulating, the mere promise to pay the metal could circulate. Unlike modern banknotes, these bills had a specified maturity but since they might be rolled over into new bills, the paper could essentially substitute for the metal. The value of the paper could even potentially exceed the amount of metal available, as long as everyone trusted everyone else.

Even without banks, it’s possible for gold reserves to represent only a fraction of the broad money circulating through an economy.

Since this thread is fairly new and uncluttered, it might be worthwhile to go through the little errors in the OP piece by piece. The main idea is correct (the conventional story of money creation from 101 classes is unnecessary and wrong) but some of the details are murky.

Although the particular link cited in the OP does fail to make a proper distinction, the conventional story can in fact make the distinction. See here for instance.

The reserve requirement can, in fact, easily restrict lending because a new loan will deplete reserves if the loan is deposited in another bank. This should be obvious.

The bank can find new reserves, of course, but the key question is whether the new reserves can be found at the right price. If replacing the reserves is too expensive, then the loan will not be profitable. It’s easy to see how reserve requirements can easily restrict lending but more on this below.

This is basically the same as the first point.

What follows is a multi-paragraph example that is broadly true for quite a long time. It’s a pretty good explanation, but I won’t copy it all.

Another problem develops near the end.

As a practical matter, the basic idea here in this paragraph is true. In the short-term, the Fed has little choice but to provide liquidity if banks are strapped.

But it’s worth emphasizing that this idea is not true theoretically, and not true in the long-term. The point is this: There is always money available. The issue is never whether banks can find reserves. Reserves are always available, and they don’t have to come from the Fed.

Suppose your local bank raised the rate on deposit accounts to 10%. Then what would people do? They would gather all the cash they could to deposit at the bank in order to take advantage of this great return. Banks could easily gain the reserves they needed not from the Fed, but from the population at large. There is over 1.3 trillion worth of cash out there in the world outside of bank vaults, and if the banks wanted that cash back, they could just pay more interest to gather it up.

The problem is never the amount of the reserves. The problem is the cost of gathering reserves. Banks will fail because the cost is too high: the interest they have to pay on reserves is too much to maintain profitability. When the Fed “raises rates”, what they are effectively doing is telling banks that liquidity will be somewhat more scarce and costly in the future: Reserves will be more expensive. And of course, if a bank cannot maintain its legally mandated reserves, the Fed will in fact charge them a penalty. A “reserves deficiency” will result in regulatory penalties.

And yes, some banks will fail because of this and because of other kinds of liquidity problems. See this link from the FDIC for case studies in how liquidity problems can kill a bank. (The FDIC is called upon to resolve most kinds of banks in the US when they fail.)

It should be obvious now that this is both silly and untrue.

The reserve requirement can be extremely important. But it’s not important in the way a Wikipedia article or an intro textbook might lead you to believe. The reserve requirement is one piece, among many, that determines the cost of gathering reserves. And quite naturally, the cost of gathering reserves is going to influence the profitability of banks.

There are relatively few of us, and MMT doesn’t offer much of interest.

7

What is the difference between “government money” and “commercial bank money” in your system other than that you have focused on “money directly issued by the government” and “money created through commercial bank transactions” as significant? Why not separate “money held by pharmacies” and “money held by ice cream trucks?” What is the purpose of your insistence on this distinction?

How do you account for the fact that the expansion of the money supply over time is a demonstrable fact?

Is your problem really with “fractional reserve banking” and do you believe that requiring 100% reserves would somehow solve whatever problem you are identifying? Or is this part of a larger theory about the Federal Reserve or the non-gold standard?

8

I agree with you about just about everything but I wanted to bring up one point: “banks and the FRB must have real assets backing up their liabilities.”

What is “real assets” is the question. For commercial banks, “real assets” are government money: currency and reserves. They’re both liabilities of the Fed.

For the Fed, on the other hand, assets are mostly Treasury bonds. I certainly consider them real assets. But the government can create as many bonds as it wants, and the Fed (another part of the government) can purchase as many as it wants. But since dollars can’t be redeemed by anything but dollars, there’s a sense in which government money is backed by either nothing at all, or by itself. I don’t think that’s anything to worry about: the government can always create a demand for dollars, simply by requiring people to pay taxes in dollars. (And banks constantly create a demand for dollars, since they require repayment of loans in dollars.) But I thought it was a point worth bringing up.

Hellestal: I agree, I think with most of your points. When I have more time/brain cells available, I’ll try to address them.

Haberdash:

All money is created either by commercial banks, or the government. It’s important, in my view, to understand the differences between those two kinds of money. For example, the Fed can create however much money it wants, whenever it wants. For the Fed, the amount it does create is purely a mater of policy. The Fed - unlike commercial banks - is not a for profit entity.

Commercial banks, on the other hand, are limited by their ability to find credit-worthy customers, and by government regulations. Furthermore, they have to be able to pay back depositors, if or when they demand their deposits.

Pharmacies and ice-cream trucks cannot create money. Since they can’t create money, they’re irrelevant to understanding how money gets created.

Commercial banks create money by lending it. The Fed creates money (mainly) by buying government debt. (A round-about way of saying “lending it to the government”.) The money supply has increased over time because both commercial banks and the Fed have increased lending over time.

I don’t think you actually read what I wrote. Which is fine; it was long and laborious. But to answer your question, I’m 100% against 100% reserves and the gold standard.

9

No. The currency and reserves just make up the 10% reserve requirement. I referred to the regulations independent of the reserve requirement, that a bank must have 100% (actually more) of its liabilities covered by assets (primarily customer debt instruments).

Similarly, FRB liabilities (including banknotes) are backed by hard assets. There is a circularity (FRB assets include promises to pay with FRB paper!) but this was the case even when the U.S. was on a gold standard.

10

I’ve read it a couple of times, and I’m not following you. What bad things can hypothetically happen relative to the concern you’re trying to communicate? What meaningful difference exists between “government-created” money and “government-sanctioned money, created by the banking system”? The dough certainly spends the same.

11

And yet no bank ever lends more money than it already has.

12

Anyone with a high tolerance for this sort of thing might try reading the recent working paper 529 from the Bank of England.

13

Yes. If by “hard assets” you’re talking about loans to customers, you’re right. I didn’t realize that’s what you meant by hard assets.

14

Every bank lends more money than it has. That’s the difference between a bank, and a guy who just lends you money.

15

Hellestal: I think the main mistake I made was saying “The reserve requirement, therefore, has no actual impact on bank lending or commercial money creation at all.”

I was getting carried away. Having said that, I think it’s not so much the reserve requirement that limits bank lending, as the amount of reserves the Fed makes available. If the Fed reduces the amount of reserves (which it would do by selling assets, such as Treasuries), there comes a point where the reduction in the supply of reserves drives up their price. Since the price of reserves is their interest rate, reducing supply drives up interest rates. When interest rates increase, more people want to deposit money, and fewer want to borrow. When fewer people are borrowing, or borrowing less, that reduces both bank lending and commercial money creation.

16

Supply and demand.

Not one or the other, but both. Central banks can supply more reserves or withdraw them, but they can also influence demand. If given the statutory power, they can implement higher reserve requirements, which will necessarily increase the banks’ demand for those reserves.

The Fed doesn’t much mess around with this. They’re more comfortable with influencing the supply of reserves, or really the future expected supply of reserves. Some central banks don’t even bother with reserve requirements at all. But in other countries, such as China, the central bank is comfortable increasing or decreasing reserve requirements as a regular policy tool. It’s arguable that capital requirements are normally more significant in banks’ lending decisions, but liquidity is a concern, too. To figure out the price (the interest rate) we need to consider both supply and demand.

17

One thing I’m not clear on based on what the OP said:

The OP seems to explain how the usual story is wrong or misleading when it comes to how commercial bank money is created through loans. But in the simpler matter of a bank customer simply making a cash deposit (as described in the “usual story” section quoted at the end of the OP) doesn’t the usual story basically hold up? The customer deposits currency into his account, no new money is created, and the bank has thereby already been given (by the account holder himself) the government money to be deposited into the reserve account.

18

Where do you get this idea?

19

Dumb question probably already answered in-thread but I don’t see it: What is MMT?

20

Modern Monetary Theory, of which I think OP is a follower. Most of its perspectives seem quite valid. I think MMT’ers see no problem running continual deficits in times of low inflation and high unemployment. I continue to hope a Doper economist will post a link to a scholarly article explaining and refuting MMT’ers prescriptions.