« Fiscal Clout and Federation Redesign | Main | Twin debt equilibria, and waiting for the sunspot »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Nick,

I realise you like thought experiments, but I think there are better ways of dealing with the “student confusion” to which you refer than the “banks buy houses” thought experiment: i.e. I prefer trying to accurately describe the REAL WORLD. So here goes.

Banks do two things: first, intermediate between borrowers and lenders, and second, create money for day to day transactions.

While it might seem that money is created when a bank lends money into existence for the benefit of a borrower, in fact the loan must be largely covered by depositors, bondholders and shareholders and the latter lot are by definition willing to deposit for relatively long periods (except to the extent than banks do maturity transformation). And money in deposit accounts, quite rightly, tends not to be counted as money (though practice varies from country to country, and varies with the term of deposit accounts).

I.e. I disagree with your claim in relation to the money apparently created when a bank lends that “That money keeps on circulating around the economy. Until the bank sells a house, which destroys the money it had previously created.” What you refer to there is not really money: it’s a loan.

Next, the charge made by banks can be divided into, 1, administration costs and 2, genuine interest (i.e. a reward for forgoing consumption).

Banks charge borrowers for both. In particular they charge genuine interest because banks have to pay interest to long term depositors, bond holders etc. In contrast, there is no reason for banks to charge genuine interest for the creation of day to day transaction money or what might be called “genuine” money because the bank does not forgo consumption when creating genuine money, and nor does anyone else.


Typo: 3rd paragraph: “the loan must be largely covered by depositors, bondholders” should have been “the loan must be largely covered by LONG TERM depositors…..”. Doh.

Is the bank truly creating new money? Doesn't the bank sooner or later have to find lenders prepared to provide the money used to buy the houses ?

In a simple model it would have to borrow first and then buy the houses later , and it could only do this if could find an interest rate where it could borrow cheaper than the ROI on the rent it got on the houses it bought.

In a model where the CB targeted interest rates then the banks could buy houses first knowing that the CB would eventually create enough money to provide sufficient deposits cover the loans. Some of this money would come from people depositing the money they received when they sold the houses to the banks, and the rest from money created by the CB by buying enough assets to make the banks books balance. The banks would buy houses (and the CB create new money) until the ROI on rents was equal to the rate they knew the CB was targeting (plus a markup for bank profit).

So the banks "create" money only to the extent that others are prepared (sooner or later) to put it in bank deposits and not spend it. It is the CB that truly creates money by buying other assets for new "outside" money.

"Banks might need higher capital ratios, because houses are normally riskier than mortgages on houses. But the liability side of banks' balance sheets could be exactly the same as now."

Let's take that to the limit. What if the capital requirement for houses is 100%?

"Maybe we should change the definition to: a "bank" is anything that creates money. What it spends that money on is immaterial."

What if the bank starts buying assets that go down in value?

"But we students of money and banking would avoid some common mistakes, like confusing the demand for money with the demand for loans. And we would see that financial intermediation has no necessary connection with money."

This is exactly why monetary policy is inefficient IMO. People demand more money but they don't demand more credit and maybe even wish to develerage. Therefore it doesn't matter if you reduce the rate on lending people don't want to borrow more which makes the broad money supply expand at insufficient pace which means growth, demand and inflation are below ideal levels. To solve this the CB needs to break the unnecessary link between lending and monetary policy. The CB can directly affect broader money interest rates or quantity in order to pursue its policy aims by effecting or placing funds into accounts of public directly.

"The one truly important and puzzling thing about banks is that people are willing to use their liabilities as a medium of exchange. But that's equally important and puzzling about any thing that creates money, whether it's a financial intermediary or not."

I'd say using demand deposits as MOE is about 1 to 1 convertibility to currency. I'd say that 1 to 1 convertibility also makes demand deposits medium of account (MOA).

"The one truly important and puzzling thing about banks is that people are willing to use their liabilities as a medium of exchange. But that's equally important and puzzling about any thing that creates money, whether it's a financial intermediary or not."

People use bank liabs as a medium of exchange because they can deposit and transact in commercial bank deposits. If People could directly deposit and transact in base money then people would use base. People are also misled to a large extent into believing that commercial bank liabs are US dollars which they are not, they are just claims on US dollars.

Nick,

When you say "money", I find it hard to tell whether you are talking about transaction accounts or about bank liabilities generally. You have told me before that you usually mean transaction accounts, as that is the medium of exchange. But the statement "the demand for houses by banks would be exactly the same as the supply of money by banks" suggests you are thinking about the totality of bank liabilities.

The problem is that money defined as transaction accounts behaves in a certain way, and money defined as total bank liabilities behaves in a way that is related but different. Conflating the two definitions actually seems to me to be one of the big sources of confusion over banks as creators of money.

Mike,

You say, “If People could directly deposit and transact in base money then people would use base.” First (a minor point) people actually do “transact” in base money when using physical cash (dollar bills).

Second (a more interesting point) I suggest people also to some extent transact in base money when doing book-keeping type money transactions (check and plastic card transactions). To illustrate….

Suppose I sell $X of government debt, and the proceeds of that sale is the only money I have. I can’t DIRECTLY transact in it, as I don’t have an account at the central bank, but to all intents and purposes I can: in effect my commercial bank would act as go-between or agent for me when I wished to pay someone some of my base money.

Moreover it would be possible to totally displace commercial bank created money with base money, which is what advocates of full reserve banking want to do. Commercial banks’ reserve ratios would need to be gradually raised to 100%, while the government/central bank machine spent freshly created base money into the economy to replace the commercial bank money being withdrawn.

Ralph: "I prefer trying to accurately describe the REAL WORLD."

I don't. I prefer trying to accurately EXPLAIN the real world. And in order to explain the real world, we need to find out which bits of the real world are important and which aren't, and which bits of the real world would change if we changed other bits of the real world.

In the real world, banks both borrow and lend. They sell their own IOUs (borrow) and buy other people's IOUs (lend). But what would change if banks only borrowed, and did not lend. They bought houses instead of other people's IOUs. Not much. Which means that banks' lending isn't very important.

"I.e. I disagree with your claim in relation to the money apparently created when a bank lends that “That money keeps on circulating around the economy. Until the bank sells a house, which destroys the money it had previously created.” What you refer to there is not really money: it’s a loan."

The money that banks create is a loan. It is an IOU, signed by the bank. It is a loan to the bank by whoever owns that IOU. Banks borrow when they create money.

Ron: let's put it this way. People must be willing to use the money created by banks. They must find it more convenient in some cases to use commercial bank money rather than only using shells, or central bank money. Commercial bank money must be able to compete with central bank money.

TMF: "Let's take that to the limit. What if the capital requirement for houses is 100%?"

Then people wouldn't be able to borrow money from the bank to help them buy a house.

TMF: "I'd say using demand deposits as MOE is about 1 to 1 convertibility to currency. I'd say that 1 to 1 convertibility also makes demand deposits medium of account (MOA)."

But you or I or any solvent person can issue IOUs that are 1 to 1 convertible into money. But nobody uses my IOUs as money.

Mike: I agree that there is something peculiar in or monetary system. Linking the supply of money with the market for loanable funds is as peculiar as linking the supply of money with the market for houses, or the market for precious metals.

"If People could directly deposit and transact in base money then people would use base."

Currency is sometimes more and sometimes less convenient than chequing accounts. Central banks don't produce chequing accounts (except for commercial banks and the government), but commercial banks are (usually) prohibited from issuing currency. They have divided up the market between them, and don't allow direct competition.

Nick Edmonds: there are companies that own houses (more usually apartment buildings) that they rent out to tenants, and they finance those purchases of houses by issuing shares and bonds (that don't circulate as money). To the extent that banks finance their purchases of houses by issuing long term non-monetary liabilities, banks are no different from those companies. (What's the correct name for such companies?)

"You say, “If People could directly deposit and transact in base money then people would use base.” First (a minor point) people actually do “transact” in base money when using physical cash (dollar bills). "

Obviously, but it is very inconvenient to transact in physical unless your talking about relatively small face to face transactions when compared to electronic pmt systems that exist for commercial bank deposits. People cant deposit their US dollars directly at a CB like commercial banks can. They just take your base if you deposit with them and give you claims on USDs in your account.

"Suppose I sell $X of government debt, and the proceeds of that sale is the only money I have. I can’t DIRECTLY transact in it, as I don’t have an account at the central bank, but to all intents and purposes I can: in effect my commercial bank would act as go-between or agent for me when I wished to pay someone some of my base money."

If people want to pay with their deposits held at the CB directly then allow it in the same way people make payments with commercial bank deposits. The central bank should be more accesible to all in providing depository services should they require them becuase the CB is the safest institution and deposits and payments are highly systemic. There is no benefit to intermediating in payments or deposits.

"Moreover it would be possible to totally displace commercial bank created money with base money, which is what advocates of full reserve banking want to do. Commercial banks’ reserve ratios would need to be gradually raised to 100%, while the government/central bank machine spent freshly created base money into the economy to replace the commercial bank money being withdrawn."

Allow people to directly transact and deposit base or commercial bank deposits. Whichever is superior will be the largest component of broad money. There is no need to be restrictive just allow healthy competition.

Nick Rowe

If the central bank allowed deposit taking and payments alongside the private banks it should be beneficial due economies of scale and scope plus the depository and payments system is systemic so people can access the safety of the CB if needed.

Ron: "So the banks "create" money only to the extent that others are prepared (sooner or later) to put it in bank deposits and not spend it. It is the CB that truly creates money by buying other assets for new "outside" money."

If the velocity of circulation of their money were infinite, neither commercial banks nor central banks would be able to create strictly positive stocks of money.

Mike: maybe. But lots of people have made similar arguments for nationalising lots of industries, and the results haven't always turned out so well. Historically, some governments have managed central banks very badly. Why would we expect them to do better with commercial banks?

TMF: "What if the bank starts buying assets that go down in value?"

Exactly the same as happens now when the bank buys assets that go down in value. Houses go up and down in value, relative to deposits; mortgages on houses go up and down in value, relative to deposits. That's why banks have capital reserves.

Nick Rowe

"Mike: maybe. But lots of people have made similar arguments for nationalising lots of industries, and the results haven't always turned out so well. Historically, some governments have managed central banks very badly. Why would we expect them to do better with commercial banks?"

Im not suggesting anything be changed with banks. All Im saying is make the depository system of the central bank accesible to anyone if they want to use it. I'm not suggesting full reserve banking or nationalized banking.


Mike: understood. But compare it to the question of whether government should get into the business of producing cars, rather than just producing roads and running bus services. (OK, not the best analogy.)

I love this post. I had to make a similar scenario in my head before I understood how QE works. With QE central banks don't buy houses but they can buy things such as MBS or corporate bonds that are a little bit more tangible (IMO) than government bonds. I figure the more aggressive QE is, the wider the selection of assets bought by central banks have to be, and the more chance they have to dip into pools of things that are tied to or backed by things that are tangible. They can even buy gold and quell the fear of those that think money should backed by more of it! :-)

These scenarios always lead me to wonder how much of the market, central banks have to replace, that is how much stuff do they have to buy, before we can reach a sensible GDP target. Also how can they choose what to buy in order to distort the free market the least possible? When QE consists of buying mostly just houses (or MBSs) aren't you favoring the housing sector over other sectors?

Nick Rowe

"Mike: understood. But compare it to the question of whether government should get into the business of producing cars, rather than just producing roads and running bus services. (OK, not the best analogy.)"

I hope I understand the question. The gov could probably produce cars relatively competitively if it wanted but arent we then loading the gov with too many responsibilities? If we overload the gov then it becomes less effective at everything becuase it cant focus and becuase of diseconomies of scale amongst other things.

But the CB already produces money and already provides deposits to some institutions. Deposit taking is a simple and systemic activity and it isnt much of a greater burden than the one it already has to extend deposit taking to the public. Also if performed by an independant central bank it wont be overburdening the gov becuase the CB is separate. The cb could be made even more independant too by making governers or commitee publicly elected.

Benoit: thanks!

(Inflammatory aside: over the years, Market Monetarists like me have taken a lot of heat from critics who accuse us of "not understanding banks" or "ignoring banks". But why aren't those critics the ones writing posts like this, asking the deeper questions like whether it really matters whether banks are or are not financial intermediaries? All they ever seem to do is fuss around with balance sheets. I accuse those same critics of not understanding money. They ignore the differences between money and all other goods: the fact that money has many markets and other goods have one. They say nothing about why people are willing to use money as a medium of exchange, and that that creates a demand for money that makes banks possible. They get confused between money and credit. How can they understand banks if they don't understand money? End of rant!)

When banks are expanding, that favours those goods that banks buy. But when banks are contracting that disfavours those same goods that banks now sell. But I am less concerned with the effect of changes in money on the markets for those particular goods. I am much more concerned about the effects of changes in the markets for those particular goods on money. Because something that affects money will affect all markets.

Mike: I think those are the sorts of questions to ask. But I'm not sure how simple the deposit-taking business really is, and whether the government would do it well or screw it up. But we are wandering too far off-topic now.

Nick,

Call them PropCos. To make that work, I have to have a way in which the public can easily switch between money and PropCo shares. After all, if I want to transfer part of my balance with the bank from a transaction account into a deposit account, it's very easy. So we could imagine that each bank has affiliated PropCo. Whenever, one of its customers wants to make that switch, the bank then sells some housing to the PropCo in exchange for shares and sells the shares to the customer. We could then say that "the demand for houses by banks would be exactly the same as the supply of money by banks" with money meaning transaction account balances. Banks' "demand for houses" here would have to specifically refer to the houses they held directly, excluding those held in PropCos.

I find that a difficult concept to relate to banks in the real world. It seems to imply that banks' demand for asset is equal to a bit they determine themselves plus a bit equal to the amount their customers want to hold in non-transaction accounts.

Nick Edmonds: couldn't PropCos just have a chequing account at the central bank to handle transfers of deposits between PropCos in exactly the way that real world banks do? Or they could keep small reserves of central bank currency, and pay them out to those wishing to withdraw their deposits, which is much the same, except less convenient, and vulnerable to muggers. PropCos are just like open-ended mutual funds, except the own houses rather than shares.

The important difference is whether or not the liabilities of PropCos, banks, whatever, are used as media of exchange.

"What should our slogan be? "Loans create deposits!"? Maybe. "Buying houses creates deposits!"? Maybe. "Creating deposits creates deposits!". Better. "Creating money creates money!". Yep, that's it."

And don't forget the ever popular, "Banks don't buy houses out of reserves."

Mark: Ha, yes! I had forgotten that one. But of course an individual bank does buy houses out of reserves, because the seller of the house is unlikely to bank with the bank that bought the house, and even if he does, the person he buys stuff from probably won't. It's only the banking system as a whole that doesn't buy houses out of reserves, and even that depends on how the central bank responds, and the public's desired Currency/Deposit ratio.

BTW, Scott Sumner said somewhere, IIRC, that central banks aren't really banks. I'm now beginning to think that commercial banks aren't essentially banks either, because it doesn't matter whether or not they are financial intermediaries, just as it doesn't really matter what the central bank holds on the asset side of its balance sheet.

"if people didn't want to hold more money, a good monetary policy would try to ensure that banks didn't want to hold more houses. Otherwise we would get boom and inflation, as people tried to spend away the excess money."

But a rational banker would always follow good monetary policy, just like rational apple farmers would always follow good apple policy. Bankers, like farmers, would be throwing away profits if they failed to provide the public with the right amount of the stuff they produce. Of course, an irrational banker might lose assets and thus cause his money to lose value, but it would be a topsy-turvy world if bankers could cause booms by behaving irrationally.

Nick, I like this post. I like to think that perhaps I partly inspired it. :D

... since you wrote "I *think* I agree with that..." after I commented on your "Banks are special..." post that I like to think of banks in aggregate paying for everything in the private non-bank world (salaries, office supplies, dividends, donuts, electricity, loans, etc) by crediting banks deposits, and that I like to likewise think of them accepting payment for everything (principal, interest, points, fees, CDs etc) by debiting bank deposits. That way you skip all the (mostly unenlightening) reserve accounting details.

On the next layer out, the CB does exactly the same thing.

Do you think your thought experiment here supports your view of MOE vs MOA and also your view that the Law of Reflux does not hold (very much) for bank deposits, and whether the hot potato effect (HPE) applies to bank deposits? Are there any testable implications of that which a talented guy like Sadowski could find evidence for in the data? I believe that you still have a disagreement with Scott Sumner and David Glasner over MOE vs MOA, the Law of Reflux and the HPE, correct? Are there testable implications to their view vs your view regarding these issues (which I see as all being related)?

Mike Sproul: suppose Martha Stewart suddenly became more popular, and people wanted to be owner-occupiers rather than tenants, so they could paint the walls in weird stripes. Profit-maximising banks would sell houses, and the money supply would fall. The question is: would banks also cut the interest rates they paid on deposits, so that the quantity of money demanded fell too by an exactly equal amount? Wouldn't that depend on the central bank's monetary policy?

Nick... and how about your disagreement with Mike Sproul? Any more to it than battling thought experiments (nothing against thought experiments... those are my favorite kind!)? Again I'm wondering is there a consequence that we would check for in the data to each view to lend support to one view over the other? What are the implications we can check for?

Tom: thanks!

This thought-experiment is designed to make it clearer that the Law of Reflux is wrong. It is very hard for people to get confused between "the quantity of houses demanded by the non-bank public always equals the stock of houses they hold" and "the stock of money held by the non-bank public always equals the stock of money they hold".

How to test between the two approaches...that's a question I keep mulling over, but don't have a clear answer to yet. A better question would be: "How much does it matter empirically that the other side is conceptually wrong?" But normally you can only measure things like that empirically when you have two logically coherent viewpoints, so you can measure whether the data is closer to one or the other. But if the other side doesn't even make logical sense...you can't do that.

Nick, I don't really buy this:

"If banks bought houses, instead of lending people the money for people to buy houses, what would be different? Not much."

There are quite a few differences:
1) A mortgage declines in value every time a payment is made. Thus, the asset for the bank shrinks as does the customers deposit. The value of the house as asset does not shrink due to payments made. In fact, in your model the rent seems to be fully counted as income for the bank unlike a mortgage payment where some part is used to pay down principal! With that the money stock declines every time a payment is made.
2) The mortgage has fixed monetary value that only declines with payments by the borrower. The value of the house is set by market conditions and is not fixed.
3) At the end of the mortgage the borrower owns the house. In your case the bank owns the house indefinitely.

I personally don't care about money but only about my combined monetary assets (plus all non-monetary assets, of course). A monetary asset is anything that is denominated in a fixed dollar amount. Just because I "lent" the money in my checking account to my bank does not mean I would not count it as my monetary asset anymore. Similar to any treasuries I may own. The combined monetary assets of the public go up when people take out loans and go down when they are being paid back or loans default. What part of those are considered "money" hugely depends on the definition you want to use but has very little application to the real world. I don't feel any difference in my wealth if I have $1000 in cash, as bank deposit, a CD or a US treasury note.

Btw. Your thought experiment is not that far off from Islamic banking as you may know:
http://en.wikipedia.org/wiki/Islamic_banking

"But if the other side doesn't even make logical sense...you can't do that." ... Ouch! :D

It sure sounds like you are saying that "the other side" here doesn't even have a logical argument... Hmmm, that's a tough one... after following SO many thought experiments between all of you going back over a year now... I've never had the impression that "the other side" was being completely illogical. Can you identify exactly where their logic breaks down?

Also, couldn't we frame the empirical test JUST in terms of your view of the Law of Reflux? I.e. can't we test the Rowe Hypothesis of "The Law of Reflux is Wrong" in terms of the data w/o regard to whether or not "the other side" is being logical or not?

Also say you were given unprecedented control over two small brand new nearly identical nations... such that you could use them for an experiment to test the Rowe Hypothesis. How could you devise an experiment to do that? I'm asking for the outlines of a thought experiment on how we could do an empirical test here I guess.

I like this thought experiment, but saying the market for houses is just one market vs. all markets for money is a gross simplification. Almost everything can be bought on credit, and in fact, total private credit market debt is 3-4x GDP so it is not a small market. So isn't credit like money because there are many markets for credit? And if so, aren't banks important? Sorry if I misunderstood the argument.

"Maybe we should change the definition to: a "bank" is anything that creates money."

And how does it do that in the Western world? By making loans which create deposits. q. e. d.

"What it spends that money on is immaterial. It could buy houses, buy financial assets backed by houses, or it could just give it away to charity, or by dropping money out of a helicopter."

Maybe we could stick with the real world? There banks only buy financial assets backed by collateral. They don't give money away for free (charity, helicopter). They also do not buy non-financial assets outright; There are regulations prohibiting that with good reason. Otherwise, banks could just buy up everything and charge the public an arm and a leg for using it. It would essentially give the bank's owners the license to print money for their own enrichment.

jt26, I think what Nick means by "MOE" in this article:

http://worthwhile.typepad.com/worthwhile_canadian_initi/2013/08/banks-and-the-medium-of-exchange-are-both-special-or-neither-special.html

Are (in part) bank deposits, which you are calling "credit." So it looks to me like he's saying exactly what you say in the last line of your comment, right there in the title of the post!

Odie, you write: "There banks only buy financial assets backed by collateral."

I don't think that's true: when the banks in aggregate buy their employee's time (pay their salaries) or pay dividends, their electric bill, interest on deposits, rent, computers and other office supplies, etc... they do so (again in aggregate) buy crediting bank deposits. They do so out of their capital, so capital constraints apply.

Tom,

But that is not money creation. They take that out of the capital their interest income earned. Their customer deposit liabilities will stay the same but their net worth will decline. If all assets of the bank will exactly equal all customer deposit liabilities it does not matter whether the bank has a balance sheet of 10 million, 100 million, or 1 billion it will have zero capital and may have to close if it does not find additional investors. That is the big difference to Nick's thought experiment. There, the bank would create $100,000 which would increase the capital/net worth of the bank and it can then use it to buy a house. In essence, the bank's owners print their own money.

Nick,

"This thought-experiment is designed to make it clearer that the Law of Reflux is wrong."

If you mean by this that people can't get rid of an excess supply of bank liabilities by repaying loans, I'd tend to agree with that. But I come back to this point that the quantity of bank liabilities is not equal to the quantity of the medium of exchange. It's one thing in a world where banks have only houses (H) as assets and transaction accounts (M) as liabilities, so H = M. But a better analogy to the real world is H = M + D, where D is the other liabilities. So knowing H doesn't tell us M; we need another equation (or maybe two more equations and another variable - an interest rate perhaps). And when we ask what, in the real world, determines the balance between M and D, it's hard to believe that the demand for medium of exchange plays no role.

Odie, you write

"But that is not money creation."

Sure it is. Every time a bank deposit is credited that's creation of "inside money.":

http://www.minneapolisfed.org/research/sr/sr374.pdf

Likewise every time bank deposits are debited that's destruction of inside money.

"They take that out of the capital their interest income earned"

Actually it doesn't matter where they take it from, interest, retained earnings, etc. It could be shares they sold to investors, or even certain kinds of bonds they sell (sub-ordinated debt for example), or profits they obtain in any form.

"Their customer deposit liabilities will stay the same..."

The deposit liabilities on the aggregated banks' BS do not stay the same. It's easiest to see this if you consider that there's just one commercial bank. They have two choices for transacting with the public: use physical cash or credit/debit bank deposits. Every time they credit a bank deposit they make inside money. Every time they debit a bank deposit they destroy inside money. Every time they use cash they use outside money. I would bet that mostly banks only buy and sell bank deposits with cash and very little else.

"There, the bank would create $100,000 which would increase the capital/net worth of the bank and it can then use it to buy a house"

Not true: that $100k would be a liability to the bank. If the house they purchase is worth $100k, they've not improved their capital position at all in the near term, although they're hoping to improve it by collecting rent in the long term. Inside money is always a liability to the bank. The aggregated banks can and do print their own money in this sense... but every time they do they add to their liabilities.

More generally every time the aggregate banks buy or sell (because that's the business they're in... buying and selling), they put into motion a potential change to their capital position. If they sell stock or sub-ordinated debt they directly add to their capital position immediately. Same goes for making a bad loan, except that they undermine their position. Same goes for paying their electric bill: it just decreases their capital... but in that case it may well be worth it. Otherwise they hope that the trades they make improves their capital position over the long run.

Banks don't hold mortgages now, for the most part. They securitize them. It makes for a better balance sheet.

In your example, I think to get a good fit, you need to have the banks securitize the houses and sell the securities. The risks gains and income would be tranched in entertaining ways. The banks would then be fulfilling their role as repackagers of risk. People who wanted to speculate in residential real estate would buy the equity tranches instead of buying houses themselves. Banks would want to get real estate off their balance sheets, because of its bad effect on their assets at risk calculations and thus their return on equity. Banks would, of course, make money as originators, servicers and trustees, as they do now.

Banks would also still make money from underwriting, trading and services. Taking deposits and making loans is only part a big modern bank's business (or even less for a few).

Nick:

If Martha Stewart became more popular and houses were bought by people instead of banks, then at the first sign of tightness in the money market, people who wanted money would start offering other goods besides houses to the banks. Banks would buy cars instead of houses. No effect on the interest rate, the quantity of money, or anything else.

Tom:

My Paper "There's No Such Thing as Fiat Money" lists 5 or 10 empirical studies of the quantity theory vs. the backing theory by Sargent, Calomiris, Bomberger and Makinen, Bruce Smith, Thomas Cunningham, Pierre Siklos, and one or two others. They all favored the backing theory. Fischer Black (of Black-Scholes fame) wrote a 'conceptual' paper favoring the backing view way back in the 1970's.

Currently, my favorite thought experiment is an-old fashioned mint. If money is tight, people bring silver to the mint to be stamped into coins. If they have too much money, people will melt coins and the coins will reflux to bullion. The market automatically provides the right amount of coins, just like it would provide the right amount of silver spoons. The Law of Reflux works just as it should.

Now make one change. Instead of the mint stamping peoples' silver into coins, the mint keeps the silver in a vault and issues paper (or electronic) tokens that circulate instead of the coins. The tokens are easier to carry and they don't wear out. The Law of Reflux works just the same as it did with coins.

Now make another change. Instead of the mint only issuing 1-ounce tokens to people who bring in 1 oz. of actual silver, the mint starts issuing 1 oz tokens to people who bring in 1-oz. WORTH of land, wheat, bonds, etc. Once again, the Law of Reflux works just the same. If people want more money they bring in more land and get more tokens. If they want less money they redeem their tokens for land.

One last change: Suspend silver convertibility of the tokens, while maintaining bond convertibility, and still accepting the tokens for taxes, loan payments, etc. Nothing important would change, except that people would wrongly conclude that 'not silver-convertible'='unbacked', and the crazy idea of fiat money would take hold.

Nick Edmonds, you write:

"If you mean by this that people can't get rid of an excess supply of bank liabilities by repaying loans, I'd tend to agree with that."

I'm a little surprised to see you write that in light of recent discussions at monetaryrealism.com. Or am I getting your position confused with that of Ramanan? Or JKH? Is there some differences between the three of you?

Also, it seems to me that loans can always be repaid, so there has to be a room for at least some reflux doesn't there?

Mike Sproul & Nick Rowe:

We've got the makings of a great smack down event here... in one corner we've got:

"...and the crazy idea of fiat money would take hold."

and in the other:

"But if the other side doesn't even make logical sense...you can't do that."

Nick, do you have anybody else in your corner? David Laidler? Bill Woolsey?

Mike Sproul... thanks for the references... I'll check them out. But in all fairness, can you think of any evidence which tends to support the null hypothesis over that of your backing theory hypothesis? Might that include the Somali "orphaned" bank notes? Something else?

Tom:

Well, in addition to Somali shillings there's bitcoin, the Iraqi Swiss Dinar, and a few others I'm probably forgetting. Those cases are, of course, extremely rare exceptions to the rule that money is always the liability of its issuer, and backed by the issuer's assets.

Possible explanations for those unbacked moneys are (1) People expect that it will be honored (i.e., backed) sometime in the future. (2) People value it as a curiosity, like baseball cards.

Tom,

If I give you a check of $1000 does that create any money? Even when our banks credit and debit our accounts respectively? I don't think so. Same with interest income earned by the bank. Let's say I make a $2000 mortgage payment. $800 are to pay down the principal, which reduces my debt liability and the bank's mortgage asset by $800 each. $1200 does not reduce my debt liability. Instead it will be income for the bank which it will use to pay e. g. one of its employees by crediting its account. No money being created are destroyed there, simply shifting funds around. For those $1200 the bank's deposit liabilities stay the same. (More maybe later)

Odie,

"If I give you a check of $1000 does that create any money?"

No, there's not NET creation of money there.

"$1200 does not reduce my debt liability."

I agree, because your mortgage and the deposit are different things. There's two separate IOUs here: your IOU to the bank (the mortgage) and their IOU to you (your deposit). Again I'm thinking just a single commercial bank, otherwise replace "bank" with "banks in aggregate."

So what do the banks in aggregate do to accept your payment for both the principal and interest? They debit your bank deposit in both cases. Or they accept cash. In the former case (debit) inside money (bank liabilities) are destroyed. In the case of you paying $800 principal the OTHER IOU is also adjusted (your IOU to the bank: your mortgage). That's the only difference. For the $800 principal payment TWO IOUs are marked down by $800, and for the $1200 interest payment just one IOU is marked down (your deposit).

The bank will have added $1200 to its equity position (which is just an abstract dollar amount representing the value of assets in excess of the value of liabilities). What they "do with it" is a totally separate issue... maybe they pay their taxes, who knows... maybe they leave it sit there as a buffer. They don't HAVE to turn around and create more money with this difference (the equity) by crediting anybody's bank deposit. They can, but they don't have to. There's no "law of the preservation of bank deposits" in play. Those are free to grow or shrink depending on what the bank decides to do.

Nick, my response to Odie is in your spam.

Odie, my longer response is in spam, but the short answer is there's no law of the preservation of bank deposits... I agree with everything your write up until "Instead it will..." to the end of your comment: you've only described a possible outcome there. There's nothing compelling the bank to turn around and credit any entity with $1200. The fact is $2000 of net inside money was destroyed. $1200 of it representing an increase in bank equity. At that point more inside money may or may not be created in the future. Nothing compels the banks in aggregate to immediately create another $2000 or $1200 or any other amount.

In the real world, banks both borrow and lend. They sell their own IOUs (borrow) and buy other people's IOUs (lend). But what would change if banks only borrowed, and did not lend?

They wold not generate an income stream out of the houses unless they sold them regularly at a profit. What you're describing is proprietary trading. And where and how would people deposit their income? Under their mattresses? How would they exchange their bank IOUs for base money? And would happen to bank balance sheets in the process?

The money that banks create is a loan. It is an IOU, signed by the bank. It is a loan to the bank by whoever owns that IOU. Banks borrow when they create money.

Banks borrow from the central bank when they make a loan. And it's the credibility of the central bank to keep the promise that makes bank IOUs near money substitutes. An IOU from a bank is a promise to pay out central bank money, not to trade it for someone else's bank IOU.

Tom,

I only said I'd "tend to" agree. ;-)

I do think reflux can occur. For example, I think QE has caused a certain amount of reflux in the repo market.

But in general, the people that hold money are not the same people that have debt, so I don't think it's that straightforward for excess money to repay loans. I think the quantity of money does change in response to demand, but with real world bank balance sheets there are many ways it can do this without loan repayment, particularly if money only means transaction accounts. I don't think reflux plays a big part in that.

I don't think this is anything different to what I have said on the MR site (although I reserve my right not to be bound by anything I've said before). I can't speak for JKH or Ramanan. My views are certainly similar to theirs on many things, but I place less importance on reflux than I believe Ramanan does.

Mike Sproul, let me ask you: if you were invited to design the monetary systems for two (or more) separate brand new Martian colonies, and you could use the opportunity to design a definitive test for your backing theory hypothesis and/or the reflux hypothesis (as applied to bank deposits or base money, etc), how would you set that up? What else would you use that opportunity to test empirically?

Oliver, you write:

"where and how would people deposit their income?"

How about this, prior to aggregated banks creation of money:

banks: A: 0, L: 0, E: 0
public: A: 100 (value of houses), L: 0, E: 100

After money creation:

banks: A: 100 (houses), L: 100 deposits, E: 0
public: A: 100 deposits, L: 0, E: 100

So their deposits are simply typed into existence by the banks, just as they are now. Now the public has 100 units of money to run their economy with. But what about after rent is due (say 10 units worth)?

banks: A: 100 (houses), L: 90 deposits, E: 10
public: A: 90 deposits, L: 0, E: 90

Now the shareholders in the bank decide they'd like to distribute 5 of dividends to themselves:

banks: A: 100 (houses), L: 95 deposits, E: 5
public: A: 95 deposits, L: 0, E: 95

etc.

By "banks" above I mean "banks in aggregate." You can assume that behind the scenes the borrow, move, and repay central bank deposits amongst themselves to settle payments. I'm also assuming reserve and capital requirements are 0 of course. Bank deposits are a medium of exchange! I've purchased four properties and not once did it involve even a penny's worth of central bank money on my part (i.e. cash, since that's the only kind I can get my hands on). I've sure that a lot of electronic bank deposits moved around behind the scenes, but that pretty much nets to zero... unless it ends up as demand deposits, in which case the central bank generally supplies 10% of that amount through repos or loans, etc.

Oliver:

"Banks borrow from the central bank when they make a loan."

That's one way to do it. The other is the bank just credits $100 to a borrower's account in exchange for a "deposit" of the borrower's IOU. The bank coins the borrower's IOU into money, without any central bank money being involved.

... that should read "electronic central bank deposits behind the scenes"

Shoot... another one caught in spam. Is there a word count threshold?

Tom & Mike

I agree with your comments. I was trying to describe how bank IOUs are themselves references to outside money in that they trade 1:1 with it but aren't money proper themselves. But that's probably not a particularly controversial statement...

Law of reflux, it is. I'm taking your side, btw..

Maybe this is a better question: Starting at tabula rasa, what is it that gives the house its initial value of 100?

Oliver:

The house is initially worth 100 oz. of silver. Then a bank issues dollars that are backed by 1 oz. worth of stuff, so now the house costs $100. Even after silver convertibility is suspended, the dollar is still backed by 1 oz worth of stuff, so the house still costs $100. If the bank lost 40 oz worth of assets, then each dollar would be worth 0.6 oz, and the house would cost $166.67.

Also, it's pointless to declare that something is not money proper. Everything is money. Once you realize that money issued by BofA is BofA's liability, and that the value of that money is determined by BofA's assets and liabilities, it becomes clear that it doesn't matter where they draw the lines for M1, N2, etc.

Tom.
I THINK what Nick is saying is banks are not special when thinking about monetary policy because the demand for credit is not the same as demand for money. Demand for credit is about the demand for goods (houses). An alternative view could be: change in demand for money is a demand for income which is the same as demand for credit *as a claim on future income*. Credit is a claim on future earnings. Banks are lending me money based on future earnings; the house as collateral just provides risk mitigation (at least for recourse states and Canada). (Just try to buy a house with 25% equity and no income!) Ditto, student loans (especially!), car loans etc.

Tom,

"The bank will have added $1200 to its equity position."

Those $1200 will go into the bank's cash account, they don't disappear. (Cash in accounting terms, not meaning actual cash as economists usually define it.) The $1200 are owned by the bank, it can do whatever its owners pleases with it: pay dividends, buy interest bearing security, pay salaries; even buy a house ;-).

If a bank wanted to buy a house outright from newly created money they would need to do the following:

1) They issue a loan of x to itself. The loan would be its liability and the deposit would go into its cash account.
2) They buy the house by transferring the deposit from its cash account to the homeowners account.
3) They collect rent payments from the person living in there by debiting the renter's account and crediting its cash account. Afterwards they debit its cash account to pay down the principal of its loan and reduce the loan liability by the same amount. They may keep some amount as earnings.

I want to see the faces of the auditors when they find such an transaction on the books. :-) If a bank can just create deposits as it pleases why does it then ever run out of capital or needs to attracts outside investors? How can a bank ever go bankrupt then?

Mike Sproul said:"Also, it's pointless to declare that something is not money proper. Everything is money. Once you realize that money issued by BofA is BofA's liability, and that the value of that money is determined by BofA's assets and liabilities, it becomes clear that it doesn't matter where they draw the lines for M1, N2, etc."

Amen. The Fed by power of the government ensures that any kind of monetary asset can ultimately be converted into base money (currency). Hence, determining in what category your "money" falls at the moment is a rather pointless endeavor.

Odie, you write:

"Those $1200 will go into the bank's cash account, they don't disappear."

actually given your example, $2000 will disappear. $2000 of medium of exchange (bank deposits). Say an individual bank writes a vendor a check for donuts... the donut shop deposits the check. If the donut shop keeps its account at that bank their deposit is credited directly. If not, then their bank credits their deposit and the donut buying bank sends reserves to the donut shop's bank. There is such a thing as an "operational account" which is just a partition of a reserve account for that purpose. That's probably what you mean by a "cash account." "Operational account" is the way a bank auditor who comments at pragcap refers to it. Buy I'm trying to get away from doing all the reserve accounting... that's why I say the banks *in aggregate* credit deposits to pay for things. The reserve accounting in aggregate is zero sum. I demonstrate the aggregate case right here:

http://worthwhile.typepad.com/worthwhile_canadian_initi/2013/09/if-banks-bought-houses.html?cid=6a00d83451688169e2019affb2f4de970d#comment-6a00d83451688169e2019affb2f4de970d

I have a number of examples on my blog. One I modeled after a John Carney article... months later I found John commenting at pragcap and asked him to check my work. He was fine with it. I also had a bank auditor that comments at pragcap check my examples, and he too was OK with them. And one other accountant as well. Plus Cullen is OK with them. They are simplified, but they show the essence of what happens. Take a look here:

http://brown-blog-5.blogspot.com/2013/03/banking-example-4-quantitative-easing.html

http://brown-blog-5.blogspot.com/2013/03/banking-example-5-bank-spends-excess.html

I'm pretty confident that those examples capture the essence of how it works, since they passed muster with my reviewers. I've avoided getting into the details of the "operational" partition of a bank's reserve account. But keep in mind, that the banks can obtain reserves to transact with by borrowing them on the interbank market. Having sufficient capital is a plus for that... but reserves themselves are not strictly required to obtain loans for more reserves. The reserve accounting is interesting, but ultimately not very important for examining what banks do in aggregate. The point of Nick's alternate reality here is to stress that banks buy and sell. That's what they do to make money (or more accurately what the do to earn net worth, or capital, etc.). If banks bought houses instead of loans Nick's example here shows how that works. When banks in aggregate make a NET credit of bank deposits, inside money (as defined in that Fed document I linked to above) is created. When banks in aggregate make a net debit of deposits inside money is destroyed. When you strip away all the reserve accounting details and look at the net results, that's what's going on.

If the aggregate banks want to buy a piece of real estate now, in our world, to build a new branch office on say, then essentially all that happens is the seller's bank deposit is credited. Again examine John Carney's article or my blog post on how that can happen and still meet capital requirements.

That's what your steps 1, 2, and 3 boil down to. There's no need for the bank to loan itself money. As long as it's capital is in good shape, it's fine. It may do some interesting book keeping, but when you strip it down to it's essentials, a bank account somewhere just gets credited... in exactly an analogous manner to how the central bank credits Fed deposits to buy things like Tsy bonds or MBS.

Odie,

Unfortunately my thread with the bank auditor "Joe" at pragcap regarding my donuts example is gone forever since that was in the old defunct "Ask Cullen." I'm sorry that's gone because I could show you the terminology he used. He seemed quite knowledgeable... but of course my purpose was to simplify down to the bear essentials, so I didn't include all his language in my posts. I recall the "operational accounts" the word "partition" (to describe how the operational account is partitioned from the rest of a bank's Fed deposit) and the word "cash and due" as an entry on the banks' balance sheets. He also described the use of "correspondence banks" in which case banks will sometimes turn to another bank to deep a deposit for them. But again... after questioning he conceded that those were details that were not necessary to get into for the purpose of understanding what happens in a big picture sort of way (not that they aren't interesting!).

I do have a thread from the new Ask-Cullen in which Joe Franzone (an accountant) answers a similar question:

http://ask-cullen.com/when-a-bank-earns-interest-on-loan-repayments-where-does-it-show-up-on-their-balance-sheet/

Again, I asked Joe Franzone to check my work, and he blessed it. I'm definitely not an expert... but my purpose is to boil down all the unnecessary details and look at what is fundamentally going on! I'm not trying to teach anybody how to be an accountant... which would be dumb, since I'm not one of those! So I may not have all the nomenclature correct, or all the steps accounted for, but I feel pretty good about the big picture. I also have a simple post exploring the difference between capital and equity based on more lost threads with Joe (Example #7). Also I explore the capital requirements example a bit more in Examples 3.1 and 3.2.

Mike Sproul said: "Also, it's pointless to declare that something is not money proper. Everything is money."

Money can be differentiated by the liquidity premium or opportunity cost and by usage context. Dollars and Euros are both absolutely liquid, but they are useful in different contexts. They are certainly both money whether here or in Europe.

Odie, you write:

"If a bank can just create deposits as it pleases why does it then ever run out of capital or needs to attracts outside investors? How can a bank ever go bankrupt then? "

The point is when a bank creates deposits it SUBTRACTS from its capital... it doesn't ADD to it. Say a bank pays it's employees and they all have deposits at the bank. It sends them checks, they deposit them (or have direct deposit) ... and cutting through any unnecessary accounting details, essentially what happens is all their deposits get credited by the amount of their paychecks. Say they get credited by $100k in total. Now the bank has $100k less equity (and perhaps $100k less capital too).

So crediting bank deposits doesn't come for free... it directly affects capital in a negative way (depending on what is purchased). If loans or houses are purchased at fair market values, then their capital is not immediately affected since they add an offsetting asset to the new liabilities (the new deposits they created). Again take a look at my example #3 the John Carney one... I had John look at all three (3, 3.1 and 3.2) but #3 is the simplest.

jt, you write:

"I THINK what Nick is saying is banks are not special when thinking about monetary policy because the demand for credit is not the same as demand for money."

Hmmm, I don't think so. Perhaps Nick will address this again to clarify, but he literally says in the post above:

"The market for houses is not the same as the market for money."

"houses" here in this post of Nick's is the analog of "loans" in the real world. Loans and deposits are two separate things... completely independent from one another after they are created. Loans are not what I think of as "credit." Loans are a legal obligation to repay a certain amount using a set schedule or formula, and they include interest obligations and perhaps even pre-payment penalties. They are not very liquid. Credit is the liquid bank deposit that the banking system in aggregate has extended in exchange for the loan. They do not have a set payment schedule: they are available on demand. That's what I think Nick means by "money" in the above quote. The ultimate deposit holder is a creditor to the bank and the bank is the depositor's debtor. Normally the deposit holder would be the car seller for example. The car buyer is a debtor to the bank and the bank (a different bank probably) is his creditor. Two different IOUs (demand deposit and car loan) two (potentially) different creditor/debtor pairs. The market for one type of IOU is not the market for the other.

Credit = money = banks deposits: it's completely liquid and is used as a medium of exchange. It is a component of M1 in the United States:

"M1: The total amount of M0 (cash/coin) outside of the private banking system plus the amount of demand deposits, travelers checks and other checkable deposits"

http://en.wikipedia.org/wiki/Money_supply#United_States

Odie, the "donut example" I refer to above is in a comment caught in spam.

Nick's post said: "TMF: "Let's take that to the limit. What if the capital requirement for houses is 100%?"

Then people wouldn't be able to borrow money from the bank to help them buy a house."

Assume a 100% capital requirement for houses and a 20% capital requirement for mortgages for houses. People could still borrow "money" to buy the house. The bank could still buy houses.

I save $200,000 in demand deposits. A new bank sells me new equity for $200,000. I believe the accountants would say the demand deposits are destroyed. Now the bank can recreate $200,000 in demand deposits to buy houses directly. However for mortgages, the bank can recreate $200,000 in demand deposits plus $800,000 in new demand deposits. Now make the capital requirement for mortgages 100%. The bank can recreate $200,000 in demand deposits to "buy" the mortgage so the borrower can spend $200,000 on the house.


And, "TMF: "I'd say using demand deposits as MOE is about 1 to 1 convertibility to currency. I'd say that 1 to 1 convertibility also makes demand deposits medium of account (MOA)."

But you or I or any solvent person can issue IOUs that are 1 to 1 convertible into money. But nobody uses my IOUs as money."

The banking system covered by the central bank most likely won't accept an you or I demand deposit. Anyone accepting a you or I demand deposit may have that demand deposit fall in value. If demand deposits stop being 1 to 1 convertible, then people will stop using them as "money" if they use them at all.

The house is initially worth 100 oz. of silver.

Just as it's worth 1000 bananas or 40 massages. Relative prices but nothing absolute. Why is the 100 oz. of silver worth a house?


Then a bank issues dollars that are backed by 1 oz. worth of stuff, so now the house costs $100.

All you're really saying to me is: assets = liabilities. But then again, my question may have been dumb.


If loans or houses are purchased at fair market values, then their capital is not immediately affected since they add an offsetting asset to the new liabilities (the new deposits they created).

So bank credit / bank purchases determine the prices on the market they are lending to / buying in. And any change in policy means that conditions of those investments (whether their own or their customer's) will be affected directly. I.e. banks cannot change the amount of money in circulation without also affecting something in the real economy. By promising to buy houses at a fixed price, when demand for houses by the non bank public decreases banks must buy them / take them off the market, and by doing so are conserving the law of house reflux.

Also interesting comment by Odie above, in that banks would have to lend to themselves against their existing capital to make such purchases. What I don't see is where added value enters the story. The base case of circuit theory, in which banks lend to firms to cover expenses during a production cycle, seems to me the more robust building block in this respect. The expenses flow back when output is bought, thus extinguishing the liability and closing the circuit. Credit finances (should finance) the addition of value to existing stuff.

Oliver, you write re: Odie's comment: "banks would have to lend to themselves"

I don't know why that would be. Maybe Odie knows more about this than I do (quite possible!) but I don't see why they'd need to lend to themselves in ANY circumstances except perhaps to fulfill some accounting requirement.

Think about it this way: would they need to lend to themselves to buy donuts, pay their electric bill, or pay their employees (say all their employees had accounts at the bank they work for)? I don't think so: ultimately the employee deposits are just credited, no matter the accounting shenanigans, which directly increases the bank's liabilities, thus reducing the bank's capital. Why would buying anything else be any different (electricity, office supplies, donuts, real estate, Tsy debt, or loans)? Sure some of those things add an offsetting asset (i.e. real estate, Tsy debt, and loans) thus equity is (mostly) preserved... and perhaps even given an opportunity to increase in the future: after all, that's the bank's business model!... buying stuff which provides a "spread" on their balance sheet to improve long term capital. Once capital is well enough in excess of requirements... the shareholders get their deposits credited too with dividends (driving capital right back down again).

The story doesn't change if we talk about the banks in aggregate and allow some employees to have deposits at other banks... now reserves will move around behind the scenes, but if reserve requirements = 0%, it makes no difference on the consolidated banks' balance sheet: some banks will be creditors and some debtors (most likely to each other, not the CB) as far as central bank deposits (i.e. reserves) go, but on a consolidated basis that all washes out.

Re: banks lending to firms: loans don't need to necessarily be repaid... credit card companies might be happy if you (and your descendents) carried a nice chunk of debt forever.

except perhaps to fulfill some accounting requirement.

That's probably the reason. Odie obviously knows more about this than both uf os. Question is, do you have a plausible theory why there should be no such requirement? The thought of banks just financing themselves instead of entreprise, seems enough reason for me to think it's a necessary accounting rule.

but on a consolidated basis that all washes out

...until someone withdraws cash from a bank at which point the whole system becomes indebted to the central bank? To me, bank deposits are always options to draw down CB money, even if that option is seldom acted upon nowadays. And CB money is that which can be used to settle all debts finally (within a currency area).

and re: loans don't need to necessarily be repaid

I agree, but as an analytical device it makes sense imo to look at closed circuits, even if they are rare birds in reality, otherwise one ends up with infinite regressions.

Come to think of it, since interest payments do not reduce the size of bank balance sheets and thus the total debt load (principal still remains), maybe the CB policy tool lever should be some sort of an amortisation rate and not an interest rate?

Nick,

why don't banks buy houses in this way?

Philippe: good question. Probably because banks' monetary liabilities are a financial instrument, and the people who are good at managing financial instruments aren't very good at doing plumbing and fixing roofs. Comparative advantage and all that. Plus you normally want to try to make the risks of your assets correlate with the risks of your liabilities, and inflation does roughly the same thing to the real value of your nominal assets and liabilities?

Tom Brown at 1.17am:

Nearly right.

"Credit" is just a silly name for IOUs. Some IOU's are used as money, but most IOUs are not used as money. (Bank of Canada and some Bank of Montreal IOUs are used as money, but nobody uses Nick Rowe IOUs as money.) The demand and supply of money is different from the demand and supply of non-monetary IOUs, and different from the demand and supply of houses. By replacing "non-monetary IOUs" with "houses" I wanted to clarify things. And you (Tom) got that point, but a load of others still don't get it. So I failed.

BTW, I now have the answer to your empirical question. But I need to think how to explain it more clearly.

Philippe: Maybe this is a better answer: because people like having themselves as landlords, and themselves as tenants. For all sorts of moral hazard/asymmetric information/transactions costs reasons. I think it gets back to the Coase question: why do firms exist? Like: why don't banks and property companies (that own apartment buildings) simply merge?

Tom,
I looked at Joe's example at PragCap. I think the confusion arises from the difference of the bank creating money or just buying assets with the money it has. In Joe's example the bank starts with $10 in equity which they lend out. However, it is their money to begin with; they can do whatever they want with it like buying donuts. Not any different as when you would lend $10 to someone. Now, Nick talked about banks creating money and buying houses. For that look at step 1 in my previous post. First, the bank creates the money x. Now the balance sheet has increased by the amount of x; new money has been created. Then the bank can spend it. Look again at Joe's post. The bank's balance sheet remains the same; the money for lending was already there to begin with. Hence, it is the bank's capital/equity. A bank has various types of assets/capital:
1. Assets in the amount of its liabilities (the big chunk). That is money that belongs someone else. They are restricted in the type of assets they can buy with it. I don't have a good source but it should be mostly financial assets. Assets the Fed would accept to provide liquidity in an event of a bank run. (Another reason why banks don't buy houses.)
2. Loan-loss reserves: A "cash" account usually part of the bank's Fed account required by regulations to absorb loan losses.
3. Equity: Anything beyond the first two categories. The bank's money. They can spend it on whatever its owner(s) pleases.
So far, those 3 categories are all previously existing money. Debits and credits within those three have no effect on the money supply.
4. "Newly created" money: The money "out of thin air" the bank creates when someone takes out a loan. It credits its customer account and accounts for the loan as asset. Afterwards, it will adjust its loan-loss reserve and check its reserve account at the Fed. The big difference is that the balance sheet has been increased by that amount from within the bank. The overall deposits within the bank system increased by that amount; new money has been created. Did that clarify it?

but a load of others still don't get it. So I failed.

That has my name all over it. But I think it's me who failed, at least in my initial reading comprehension of your post. It was late, yesterday...

Odie:

The big difference is that the balance sheet has been increased by that amount from within the bank.

And this is, I think, the decisive point. I hear Nick and Tom saying that house purchases by banks would increase bank balance sheets in the same way that loans do. And I think you're right, they don't and can't. But I'm going to have to rely on you for the moment to explain why and what this has to do with the nature of money :-).

Oliver: "I hear Nick and Tom saying that house purchases by banks would increase bank balance sheets in the same way that loans do."

You hear me right. The bank just credits the house seller's chequing account, or writes the house seller a cheque, drawn on the bank, which the house seller deposits in his chequing account, in exactly the same way as it would if the bank had made a loan.

Now why does Odie think that's wrong (if he does think it's wrong)?

Nick, (Oliver),

I get that in your model the bank would buy the house with newly created money but that is currently not possible for banks as we can see empirically. There is a simple reason: regulations. Imagine a bank's depositors would demand currency/cash out their accounts. The bank does comply with that by selling assets to the Fed who supplies the bank with currency. For your picture to work the Fed would need to accept houses as assets in order to provide banks with liquidity. It does not. We can discuss whether that makes sense or not but that would be an academic argument, not really something that applies to our current monetary system.

Second, to create new money the bank relies on customer demand for credit. It is prohibited to sign its own loan papers. Imagine you would work at a bank, you would fill out your own loan application, sign it, credit your account and spend the money on whatever you pleases. Again, you can't. You need two parties for issuing a loan. One that needs the money and one that underwrites it after checking the creditworthiness of the other party. There needs to be checks and balances. The economy decides how much deposits it desires and the banks how much risk they want to take.

Btw. My Mom deals with real estate for banks out of bankruptcies/foreclosures. I can tell you, banks HATE being landlords. ;-)

Just found this:

According to monetarism, money is indeed a positive asset allowing barter to be split into a sale and a subsequent purchase. Money is thus conceived of as a stock and the money supply is seen as the quantity of this stock determined, directly or indirectly, by monetary authorities. Yet, if money is identified with a positive asset, then banks are bound to benefit from the extraordinary power to create riches out of nothing, as it were, which is plain nonsense. In reality, banks act as monetary intermediaries, which means that money is issued as a flow any time banks carry out a payment on behalf of their clients. Every payment is a tripolar transaction involving a bank and two of its clients, in which each of the three agents involved is simultaneously a purchaser and a seller on the labour, commodity, and financial markets. As a purely numerical form money never enters a net sale or a net purchase and must therefore be clearly distinguished from bank deposits, net assets and liabilities entered as stocks in the bank’s balance sheet and that can only result from the association between money proper and real output established by production. Knapp’s idea that money is essentially state money is thus contradicted by the fact that, like any other economic agent, the state cannot finance its spending through money creation, that is, by issuing its own acknowledgment of debt. In a logical (as opposed to pathological) system, public spending is constrained by the amount of income the government can obtain through taxation, private loans, and the sale of public goods and services, which simply means that, again like any other agent, the state can finance its purchases only by simultaneous and equivalent sales on the commodity and financial markets. What lies behind the confusion between money and income is the concept of credit money and the wrong belief that when banks create money they grant a positive credit to the economy...

Odie: "The bank does comply with that by selling assets to the Fed who supplies the bank with currency. For your picture to work the Fed would need to accept houses as assets in order to provide banks with liquidity."

Suppose I get a loan from the Bank of Montreal (I sell the Band of Montreal an IOU with my signature on it). The Bank of Canada does not (normally) buy Nick Rowe IOUs from the Bank of Montreal, any more than it (normally) buys houses from the Bank of Montreal. Instead, the Bank of Canada will look at the Bank of Montreal's assets, which include Nick Rowe IOUs plus houses, when it lends to the Bank of Montreal.

"You need two parties for issuing a loan. One that needs the money and one that underwrites it after checking the creditworthiness of the other party."

You also need two parties for buying a house. The house seller must agree, and the bank must agree, and will probably want to get the house inspected before it buys it.

Nick, might it also have something to do with what banks are permitted to do by law? Recently there was a lot of discussion about how rules had been relaxed in the 2000s to allow banks to do things they previously hadn't been allowed to, such as own physical commodities and stockpile them in warehouses.

http://economix.blogs.nytimes.com/2013/08/08/getting-big-banks-out-of-the-commodities-business/?_r=0

I wonder if there are rules limiting the ability of banks to buy up the whole real estate market?

Actually I saw a 'BNP Paribas Real Estate' sign on a massive new office development the other day, so maybe banks and property companies are already merging:

http://www.realestate.bnpparibas.com/bnppre/en/home-cfo4_12097.html

Oliver:

"Knapp’s idea that money is essentially state money is thus contradicted by the fact that, like any other economic agent, the state cannot finance its spending through money creation, that is, by issuing its own acknowledgment of debt."

Check out my paper on American colonial currency. In 1690, for example, Massachusetts printed 100 paper shillings and paid soldiers with them. The shillings were backed by Mass' promise to accept the paper shillings at par with silver shillings in payment of taxes. In effect, paper shillings were backed by the government's assets, mainly 'taxes receivable'.

Philippe: Maybe. I don't know. But it is true that for some reason, banks really don't like owning houses, as Odie's mother says. I bought my house from a bank. I put in an unconditional offer with an asap closing date. The house was dirty, lawn uncut, light fixtures missing, and it was sold as is (no guarantee), so it showed really badly. But I figured there was nothing wrong with it a few days' work and a few hundred dollars wouldn't fix. I got a very good price because the bank desperately wanted it sold quickly, and few buyers would look at it. (That was well before the recent recession, BTW.)

So housing units would be the unit of account.

And that unit of account is an elastic unit of account -- in stretches and shrinks depending on the percentages put into that type of long term production good vs short term production goods producing alternative outputs.

And it is also one of changing liquidity -- houses will be more or less liquid depending on changing expectations, the unexpected discovery of real and changing resource scarcities, and changing optimism or pessimism.

Greg: "So housing units would be the unit of account."

That doesn't follow. A bank's monetary liabilities are normally defined in terms of the unit of account, simply because separation of MOE and MOA is usually inconvenient. But there is no reason its assets should be.

Odie, did you also read the part where Joe wrote to me "yes your examples are correct?" Take a look, especially at my Example #5, or the one I write out here with the houses worth 100 (above).

Odie, you write:

"3. Equity: Anything beyond the first two categories. The bank's money. They can spend it on whatever its owner(s) pleases."

Equity is an abstraction: it's the dollar value of the value of the assets in excess of the dollar value of it's liabilities. There are CAMELS requirements, a capital adequecy ratio (CAR), and reserve requirements... yes I'm away are of all that. But when the aggregated banks make a net increase in the total dollar value of the deposits they hold as liabilities, money is created in the amount of M1.

Again, imagine you are a bank employee, you have your paycheck directly deposited in your bank (which is the same one you work for). No other transactions happen in the world over that hour. What do you imagine just happened? Did somebody else lose a deposit? No: Only yours was credited.

Think of it this way, the banks assets might be all Tsy bonds and they have $100 worth. Their liabilities might be all loans of reserves that they took out from other banks or the Fed: they might have $50 of those. How much equity does the bank have? $50. $50 of "demand deposits" or "reserve deposits" or "cash?" No. Just an abstract value: $50 of assets in excess of liabilities.

Can they afford to pay you $10 for the work you did for them? Absolutely! They simply credit your checking account by typing it in, and voila! You've got $10. Now you can write a check to the grocer who also banks at that bank and your account is debited and his is credited. You and the grocer can't destroy or create bank deposits (it's "outside money" to both of you) but the bank can. The bank just created $10 of M1 out of thin air. Was it free? Absolutely NOT! The bank also just lost $10 of equity: now they've only got $40 equity. What happens if you want to take $10 in cash? Now they're on the hook to find the cash for you. Perhaps they can sell their Tsy to another bank for $100, and then buy $10 of cash from the Fed or another bank. Perhaps they have a $10 Tsy they can sell instead. They'll do what they need to. Of course they usually keep enough on hand so they don't have to do it after the fact!

Equity isn't "money": it's assets in excess of liabilities. The bank is free to turn some of their equity into money (their liability).

Nick,

You say:"Suppose I get a loan from the Bank of Montreal (I sell the Band of Montreal an IOU with my signature on it). The Bank of Canada does not (normally) buy Nick Rowe IOUs from the Bank of Montreal, any more than it (normally) buys houses from the Bank of Montreal. Instead, the Bank of Canada will look at the Bank of Montreal's assets, which include Nick Rowe IOUs plus houses, when it lends to the Bank of Montreal."

That is a better way of putting it than what I said. Essentially, the Central Bank will lend any bank money for an acceptable collateral. The Fed lists the ones for the discount window here: http://www.federalreserve.gov/newsevents/reform_discount_window.htm You see, those are all financial assets for the simple reason that the value of a performing asset is known at any point. Not so for a non-financial asset like a house. That is the tricky part about accounting to decide what is the current mark-to-market value of a non-financial asset.

Ok, let's have the discussion why the banks do not create money to buy non-monetary assets instead of issuing loans. In that scenario, for the private sector to get money into existence they would need to sell something to the bank. When you want to start a zero the bank will need to pay the contractors to build the house before it can rent it out. Not stopping at houses, people can sell their labor by working for the bank or produce a good that the bank wants to buy. This would in your model be the only way money gets into the hands of consumers and businesses. If someone wanted to start a business he would not take a out a loan but "sell" the idea and then work for the bank. Since all money created by banks is backed by the Central Bank either the government (or the bank's owners) would ultimately control most of the productive capacity of the private sector. Does that ring a bell? ;-)

Tom,

I looked at example 5. Where is money created there by issuing a loan? Bank A starts with $100 in equity; the rest are a lot of debits and credits. Can you let Bank A start at $0 and go from there?

Tom,

You said:"Again, imagine you are a bank employee, you have your paycheck directly deposited in your bank (which is the same one you work for). No other transactions happen in the world over that hour. What do you imagine just happened? Did somebody else lose a deposit? No: Only yours was credited."

Of course, the bank lost a deposit. They took that out of their "operating account" (or "cash" account when you look at a balance sheet). Again, banks don't create money to pay their employees. They just debit and credit different accounts.

You say:"Think of it this way, the banks assets might be all Tsy bonds and they have $100 worth. Their liabilities might be all loans of reserves that they took out from other banks or the Fed: they might have $50 of those. How much equity does the bank have? $50. $50 of "demand deposits" or "reserve deposits" or "cash?" No. Just an abstract value: $50 of assets in excess of liabilities. Can they afford to pay you $10 for the work you did for them? Absolutely!"

Absolutely not! That bank would be illiquid and could not pay its employees! Luckily, it has treasuries which it can sell at any time to receive a deposit. And since the bank has excess capital it can use some of the proceeds to pay its employees.

I don't get hung up with the questions of what is M0, M1, M2 and so on or "inside money" versus "outside money". I look at the balances and transactions to determine how much monetary asset has been created or destroyed. Btw. if I wanted I could easily destroy M0. I just have to light up a dollar bill.

Odie, sure. In fact that's what Example #3 does: the bank obtains $10 in "retained earnings" starting from a blank balance sheet. Again that follows almost exactly the John Carney write up I link to there (plus he later reviewed my post and blessed it).

But let's repeat it here (my numbers won't be close to realistic... I'm just trying to demonstrate a point):

bank A: A: 0, L: 0, E: 0
person x: A: 0, L: 0, E: 0

Say x takes a loan from A for 100 and agrees to pay a 20 loan origination fee (or points) which creates a grand total of 80 units of money in the world:

bank A: A: 100 loan, L: 80 deposit, E: 20
person x: A: 80 deposit, L: 100, E: -20

Now A pays x 10 units for some services to the bank (maybe he's their gardener), which results in 90 units of money in the world:

bank A: A: 100 loan, L: 90 deposit, E: 10
person x: A: 90 deposit, L: 100, E: -10

I'm assuming a 0% reserve requirement. You can assume a 10% capital adequacy ratio (CAR) requirement, which the bank will still just be able to meet after paying x (assuming the loan to x is very risky: i.e. it has a risk weighting of 1).

So in terms of total money in this world it went from 0 to 80 to 90. 0 to 80 from a loan (from the bank buying a loan) and 80 to 90 from paying for gardening services. The bank equity is an abstraction: just the dollar amount of assets in excess of liabilities. Assets may or may not be comprised of some "money" component such as electronic Fed deposits or vault cash. In this case neither is involved.

... and "retained earnings" through points or fees (or eventually through collecting interest) is only one option: the bank can also sell sub-ordinated debt (which doesn't affect equity, but does affect capital the way it's calculated... there's a difference!), or it could sell shares of stock. (again I appear to have a post in spam... so it should come out eventually). I take a look at capital vs equity in my Example #7. That was based on a thread with Joe the bank auditor (different than Joe Franzone), but I don't know that he ever reviewed it (so a word of caution there).

Odie, I disagree with this part of your post starting here:

"Absolutely not!"

if ALL the banks depositors came in and wanted cash there'd be a problem because they don't keep that much cash on hand. Does that mean they're illiquid? I don't think so. They don't have to have cash on hand to extend credit. They don't even have to have that many electronic reserves on hand. This is easiest to see in a 0% RR country like Canada. But even in the US they just need to have 10% of the DDs on hand, not 100%. I'm not saying that even in Canada banks won't keep some reserves on hand as a buffer... but they certainly don't have enough to cover all their deposits. That would make a bank's job of creating a favorable spread on it's balance sheet more difficult. They just have to be able to obtain the cash or reserves if needed. Reserves are easy since they are electronic. Read about the capital adequacy ratio (CAR). Read John Carney's article.

Again, I'm no expert, but I did have folks that ought to know look over those examples. I trust that I've basically got it correct. You have not convinced me otherwise.

Regarding burning a dollar bill: yes, I agree that is a good example. But making them is another matter! Also, try destroying an electronic bank deposit w/o withdrawing any in cash and burning it. Or making one for that matter. Pretty much you are limited to withdrawing coins and notes and physically destroying them in terms of your ability to create and destroy money, which from your point of view is "outside" money (i.e. which includes M1 and bank money). In contrast, you can go hog wild writing IOUs... creating and destroying your own "inside money." Banks can do likewise with their inside money as long as they meet the RRs, the CAR and the CAMELS.... and satisfy their shareholders.

Tom Brown, does the first part of my 3:02 a.m. post about capital requirements sound good?

Tom:
You said: credit = money
I meant: credit = loans = non-monetary IOUs
(otherwise I wouldn't have used "credit" and "money" in the same sentence)

So again: with credit->loans
I THINK what Nick is saying is banks are not special when thinking about monetary policy because the demand for loans is not the same as demand for money. Demand for loans is about the demand for goods (houses). An alternative view could be: change in demand for money is a demand for income which is the same as demand for loans *as a claim on future income*. Loans is a claim on future earnings. Banks are lending me money based on future earnings; the house as collateral just provides risk mitigation (at least for recourse states and Canada). (Just try to buy a house with 25% equity and no income!) Ditto, student loans (especially!), car loans etc.

Tom Brown, does the first part of my 3 : 02 post about capital requirements sound good?

jt26... OK, that makes what you're saying clearer. Nick clearly thinks of credit more like I do though I think (see his comments above!). But I still don't think Nick is saying that banks are not special, even in a monetary sense. They're probably not as "special" to Nick as to a PKEer, but they do have some "specialness" at least in the short term or in terms of reducing frictions. He says as much in that other post of his that I linked to: the "Banks are special..." one. And he says as much in the comments to Sumner's recent post on this:

http://www.themoneyillusion.com/?p=23908

I'll try to digest the rest of what you're saying here.

Too Much Fed: Sorry, I don't see it. Perhaps it's in spam.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad