John Maynard Keynes famously wrote that: "Practical men, who believe themselves to be quite exempt from any intellectual influence, are usually the slaves of some defunct economist." A modern example of that dictum, relevant to the economy, policy, and markets, is the widespread view that people can "spend" their money, as if money represented a pool that is just waiting to "flow into" spending. Because such a thing cannot occur and therefore has not occurred, the point is usually made in reverse: people currently are not "spending" their money–rather they are "parking" their money at the bank or leaving it "idle." But that they might spend it in the future is a lurking risk and a reason to be cautious about the central bank engaging in aggressive quantitative easing (QE).
To see what is wrong with that standard textbook view, we need to consider the following fundamental accounting identity:
M = M
The amount of money that people hold must equal the amount of money created by the banking system. You can't see "spending" anywhere in that fundamental accounting identity, can you? Therefore, people do not "spend" money!
A key distinction to bear in mind is between individual people and people in aggregate. Neither individual people nor people as a whole can "spend" money, but individual people can and do offload their money (particularly excess money) by lending it to other people or by buying goods and assets; but people in aggregate cannot do this–in such cases, the money that leaves one person's balance sheet just pops up on another person's balance sheet, remaining on the banking system's balance sheet all the while.
Therefore, people cannot and do not "spend" money. This explains why creating money does not work to increase spending. Except, maybe, via obscure indirect mechanisms.
The above is total rubbish, of course. It is also heavily plagiarised, from an article by Paul Sheard (pdf) (HT David Andolfatto). Basically, I just changed his "banks" to "people", his "lend" to "spend", and his "reserves" to "money".
Now let me be sensible:
We define "excess money" as the actual stock of money that people hold minus the stock of money they desire to hold (given prices, income, interest rates etc.).
If an individual person has excess money, he can and will get rid of it, by spending it or by lending it. (And "lending" means "buying an IOU from someone", so it's the same as spending.) But that just means another individual person now holds it.
If the banking system creates an excess supply of money, and holds that stock of money constant, each individual person can get rid of it, but people in aggregate cannot get rid of it. It just keeps circulating back to them, as quickly as they spend it. But their individual attempts to get rid of it are what create the increased demand for goods, which will ultimately raise the prices of goods above what they would otherwise have been. The fact that people cannot in aggregate get rid of the excess money is a central part of the standard story of why excess money raises the demand for goods and the prices of goods. If they could get rid of it by spending it, the excess money would have at most only a temporary effect.
It's exactly the same with excess reserves. But here we need to be careful about how we define "excess reserves". The economically relevant definition is "actual stock of reserves minus desired stock of reserves". We should not define "excess reserves" as "actual stock of reserves minus legally rquired stock of reserves". The former definition works for economists in all countries, regardless of whether or not there are legal reserve requirements (Canada has none). The latter definition is only good for US lawyers.
It makes absolutely no difference whether banks make loans in the form of currency or in the form of creating demand deposits. An individual bank that makes a loan of $100 by creating a deposit of $100 will lose $100 of reserves to a second bank when the borrower spends that $100 on a bike, and the bike seller deposits the cheque in that second bank. If the bike seller will only accept currency, so the first bank swaps $100 in reserves for $100 in currency, then lends $100 currency to the borrower, the loss in reserves is immediate, rather than delayed by a day or two. But the end result is exactly the same.
Let's cut to the goddamn chase: banks lend reserves.
And the fact that banks cannot in aggregate get rid of the excess reserves is a central part of the standard textbook story of why excess reserves raise the stock of money, which creates an excess supply of money, which raises the demand for goods and the prices of goods. If banks in aggregate could get rid of reserves by lending them, the excess reserves would have at most only a temporary effect.
(I could do another post on the "needs of trade" fallacy (the idea that "interest-rate-targeting central banks supply whatever reserves are needed") in that paper, but I have already done several related posts, like this one. So I think I will drive to Wawa instead.)
Worthwhile Canadian Initiative 
@TMF:
Majromax: “If the banking system has no excess reserves (over desired reserves), then your loan repayment will simply result in another loan being issued to the next marginal borrower, enticed by a slightly lower interest rate versus the counterfactual of keeping your loan outstanding.”
That is an empirical question, isn’t it?
@Min:
Addendum to the above: if it’s not clear from context, I think this long-term disequilibrium is applicable to the United States.
As I understand it, the Fed is paying 0.25% interest on reserves, yet the short-term rates are at 0.01% to 0.11% for maturities as far as a year out (and 0.05%-0.1% for 1-3 month commercial paper). That more than satisfies my non-clearing requirement that the banks be unable to issue profitable loans, since the risk-free rates they can achieve are below the IOR rate.
I don’t think that is applicable in Canada, however. Three-month Canadian treasury bills are a bit above 0.9%, whereas the deposit rate is 0.75%. All other things being equal, Canadian banks with excess reserves would be able to hold treasury bills instead to earn a profit, ergo they’re probably not holding on to significant excess reserves.
Majromax: “If the banking system has no excess reserves (over desired reserves), then your loan repayment will simply result in another loan being issued to the next marginal borrower, enticed by a slightly lower interest rate versus the counterfactual of keeping your loan outstanding.”
Moi: “That is an empirical question, isn’t it?”
Majromax: “Only insofar as markets are empirical. My statement about the marginal loan is just a longer-winded way of saying that the loan market clears”
It sounds like you are asserting that your statement has been empirically proven. (Bearing in mind that empirical proof is different from logical or mathematical proof.) But I find it hard to come up with an operational definition of a market clearing when nothing in the market exists before the trade. In real life there are always things that do not happen but could.
Anyway, I will not press you for empirical proof. We are both mere commenters, after all. 🙂
Analogy of what is being critiqued:
….”To see what is wrong with that standard textbook view, we need to consider the following fundamental accounting identity:
….”M = M
….”The amount of money that people hold must equal the amount of money created by the banking system. You can’t see “spending” anywhere in that fundamental accounting identity, can you? Therefore, people do not “spend” money!”
Now, the equation should really be M1 = M0, or the like, because the point is that people do not create money by spending it, so that the resulting amount of money, M1 is equal to the previous amount of money, M0. Now, OC, the BS comes with redefining the verb, “spend”. The point is that banks create money, not people.
There is another problem, too, however. A couple of nights ago I went out and bought dinner for $20, paying, not with money, but with a credit card. The dinner was paid for by money created by the bank in response to my card charge. So in that case, M1 > M0. Accounting identity indeed! We can also view that as my creating money. I did not do so myself, but because of my agreement with the bank issuing the credit card, they did it at my behest.
Now, in the analogy:
….’Basically, I just changed his “banks” to “people”, his “lend” to “spend”, and his “reserves” to “money”.’
OK, R1 = R0, because banks do not create reserves. But what about that analogy between spending and lending? People do not create the money they spend, unless we count using credit cards and such.
….”(And “lending” means “buying an IOU from someone”, so it’s the same as spending.)”
Whoa, Nellie! Thet’s true if people make loans, because people do not create the money that they spend (as long as M1 = M0). But the analogy does not carry over to banks, because they do create the money that they lend, even if R1 = R0. (Under our current system, OC.) Lending by banks does not simply mean buying an IOU, it also means creating the money to do so. This is uncontroversial, right?
….”Let’s cut to the goddamn chase: banks lend reserves.”
Banks do not create reserves themselves, right? R1 = R0. Well, if banks simply lend reserves, why in hell do they create money?
“banks lend reserves”
Needs to be defined.
Try this scenario.
Run the model with only one commercial bank. Are reserves being lent?
Run the model with only one person. Is money being spent?
Philippe, there will be big differences between a one person economy and a one commercial bank economy.
Majromax said: “Then we’re at the zero lower bound, aren’t we?”
What if a bank decided to do something like “QE”?
Nick, you’ve written a lot of good posts on this subject, this is another one. A question I still have (that I hope you’ll address in future posts) is this: Your line of reasoning (with which I concur) is used to justify models assuming that saving drives investment. For example, the Solow model is frequently used to “demonstrate” what would happen if the government implements policies that encourage more household saving. The “answer” is that this would make us all better off in the long run (assuming the current capital stock is below the “golden rule” level) because it would increase investment. Do decisions by economic agents about how much to “save” (as opposed to decisions about how much to borrow for investment purposes) matter all that much if the central bank is targeting inflation? Economics students are repeatedly told that more “saving” leads to lower interest rates and thus more investment and a larger capital stock. Is this so? (I’m referring to long-term trends in saving as opposed to short-term fluctuations). Thank you.
Would be better to tax the rich and redistribute the wealth when we have less than full employment…. but the rich won’t let that happen…. they control tax laws.. they hoard and hide….and so we are left with printing money…….
You are saying that we are facing the risk that money in the hands of the wealthy will become a “high multiplier” place?
We can only wish
Mr. Rowe, much of you write here is correct — and very clear. But there are several little points which are wrong. Others have already addressed several of them better than I could. I apologize for the length of this.
One point is key for understanding what’s going on.
You write:
“(And “lending” means “buying an IOU from someone”, so it’s the same as spending.) ”
Lending does mean “buying an IOU from someone”.
However, IOUs are a substitute for money. If an IOU is tradeable and liquid, it actually is money. So this is not the same as spending.
This is essentially how money is created by private banks and other entities. You know this. These IOUs are bank-created money (or, alternativey, they are causing a drop in demand for government money). This means that even with constant “government money”, the effective money stock isn’t constant (or, alternatively, the demand is affected by the creation of IOUs).
Perhaps your January explanation was the better one, since it seems clearer:
http://worthwhile.typepad.com/worthwhile_canadian_initi/2013/01/two-extreme-fiscalmonetary-worlds.html
“…The government in the second, monetary world, has two degrees of freedom. It can choose how many liabilities to sell, and it can choose the rate of interest it pays on those liabilities.
“The government in the real world is a combination of those two extremes, and it can choose what combination it is. That gives it three degrees of freedom. It can choose how many liabilities to sell. It can choose what combination of liabilities to sell. And it can choose what rate of interest to pay on its monetary liabilities. Because money is different from all other assets, financial or otherwise.”
I do have one dispute: I deny that there is an actual third degree of freedom. Effectively, I believe we are in “monetary world” in all the countries which haven’t “dollarized” or joined the Euro. In short, T-bills or gilts are treated as money. Very large denomination money, but money.
Forget the sort of bonds which are not treated as money for now; they are not relevant.
There is red money (bank-issued money) and green money (government-issued money). The government has two degrees of freedom: quantity of money and interest rate given on money.
The commercial banks also have two degrees of freedom. They can adjust the rate of interest they offer to pay on “red money”, which changes people’s preference between “red money” and “green money”. The banks cannot usually offer a lower rate of interest on red money than the government offers on green money (but sometimes they can, if they can claim that depositing in their bank is “safer” than holding piles of cash under your mattress). People’s preference between “red money” and “green money” is not just determined by the interest rate, it’s also determined by how irresponsible the recent behavior of the banks has been.
Nobody refuses “red money” — checks, for example, are “red money”, as are debit cards. Nobody stops to check whether the bank issuing the check is solvent (well, not anymore, they used to).
But the banks can also determine how much “red money” to offer! This gives them a second degree of freedom. It is this degree of freedom which we generally attempt to regulate, generally through abstruse and baroque methods involving “reserve requirements”, “capital requirements” and so on and so forth. For the sake of my argument, imagine a country where the bank regulators are out to lunch and routinely rubberstamp anything including completely fraudulent accounting. (We’ll call it the “US”.) Here, it is clear that the banks have this second degree of freedom.
The amount of “red money” which is actually supplied depends also on the demand for it, which is also determined by the risk taking profile of the (relatively) wealthy.
Thank you for helping me clarify my thoughts; the January entry is particularly clear.
I hope this helps you clarify yours; the main oversight in most of the stuff I’ve read by you has always been bank-created money (red money) and the weird effects it has on the economy. Since the role of bank-created money — and its demonetization as people suddenly decided that money market funds weren’t a medium of exchange (they were on a Monday, they weren’t on a Tuesday) — is essential to the understanding the crash of 2008, I think incorporating this concept would probably assist your work.
As others have noted, it is an empirical fact that bank loan departments — at least in the US, in the 21st century — do not pay attention to reserve or capital levels when they make loans. Lending supply volume is not dependent on reserves in any sense, and banks strive to make it as large as possible given the bank’s chosen interest rate. If the loans they issue cause the bank to have a shortage of government money, the bank simply knocks on the discount window and the Federal Reserve shovels some more government money into it. This is the entire discount window policy.
(Banks can choose their interest rate on loans in thee short term, determined by desired profit margins. In the medium term it is determined by competition with other banks — or cartel arrangements.)
If borrowing at the discount window causes the bank to become insolvent, well, frankly in my hypothetical NO-REGULATOR world of the US, this has no effect at all; the bank continues operating and getting free money shovelled at it; the bank-issued money (deposits, commercial paper, etc.) is still treated as good money, etc. I believe we’re watching this happen right now in the US.
Only if there is a threat of shutdown by the regulator (the FDIC or one of the other ones in the US) is there a run on the bank and demonetization of its deposits.
This gets back to Min’s comment about the “bank in Rhode Island that issued some $600,000 in bank notes with 7 bits of a Spanish Dollar ($0.875) in the vault.”
We also have shadow banks, like the money market funds. They don’t have access to the discount window, so they can be demonetized much much faster when they run out of government money and par convertibility becomes impossible. They depend on a supply of reserves to avoid bank runs.
In any case, banks issue loans without regard to reserves. They deliver on the loans partly by actually printing banknotes (in the 19th century… or some Scottish banks), and partly by printing money in the form of checks (“here’s your loan, the money is in your deposit account”). When forced to, they issue government currency; if they run out of it, they go to the Fed to get more currency, which the Fed promises to give to them. In order to pay less in interest to the Fed, they convince depositors to give them government money in exchange for “red money” (bank deposits).
Market Fiscalist describes this view, and it is correct:
“2. The amount of lending in the system is determined by the CB target interest rate and not the amount of reserves. The CB will adjust the money supply to make sure banks can get the reserves they need to back current lending.”
The second sentence is simply an empirical fact about central bank policy. It is debatable whether the amount of lending in the system is determined by the CB target interest rate, but as long as there’s a discount window, it is certainly not determined by the amount of reserves.
There are two different types of banks: those which are backed by a discount window and those which aren’t (like the money market funds). They have different dynamics and different behavior, because the former don’t need reserves and the latter do. Both types can and do print money.
I wouldn’t mind seeing a really coherent stock-flow model of the banking system, including both types of banks, and separate elements for rich people and poor people. I’ve never seen one. The key number which matters for the macroeconomy is the flow of money available to the 99%, because that’s what determines productive activity, purchase of stuff like food and cars. This is determined by a lot of things including willingness to borrow by the 99% and creditworthiness of the 99%.
At the moment we have high stocks of money held by the 1% and very low stocks accessible to the 99%. The maldistribution makes a lot of things happen which seem impossible in models which assume a single uniform representative individual.
Min said: “There is another problem, too, however. A couple of nights ago I went out and bought dinner for $20, paying, not with money, but with a credit card. The dinner was paid for by money created by the bank in response to my card charge. So in that case, M1 > M0. Accounting identity indeed! We can also view that as my creating money. I did not do so myself, but because of my agreement with the bank issuing the credit card, they did it at my behest.”
Min, do you think demand deposits are both medium of account (MOA) and medium of exchange (MOE)?
Min, what do you think of this description of your transaction?
Min balance sheet: create a new bond that is both your asset and your liability
Bank balance sheet: create new demand deposits that are both the bank’s asset and the bank’s liability
Asset swap the bond for the demand deposits to your credit card account (similar to a checking account). The demand deposits become your asset. The bond becomes the bank’s asset.
Asset swap the demand deposits for the food.
Too Much Fed: “Min, do you think demand deposits are both medium of account (MOA) and medium of exchange (MOE)?”
First, I think that money is a fuzzy concept, in the technical sense. That’s why we have M0, M3, etc.
IIUC, medium of account is anything denominated in terms of the standard unit of account. Last time I looked, my demand deposits were denominated in US Dollars.
“IIUC, medium of account is anything denominated in terms of the standard unit of account.”
Min, I agree with this part at this link.
http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/10/medium-of-account-vs-medium-of-exchange/comments/page/1/#comments
“[Update: just to clarify terminology: in my model, gold is the medium of account; and (say) an ounce of gold is the unit of account.]”
Nick talks somewhat about fixing prices involving MOA and MOE, but I am not sure I agree with all there.
O/T: Any macro types want to take up this challenge?
http://informationtransfereconomics.blogspot.com/2014/07/a-challenge-to-macroeconomists.html
Nick,
No-one has ever said that lending does not result in reserve movements. But in my work I separate out lending and settlement, whereas you conflate the two. Lending itself DOES NOT result in reserve movements. It is the drawdown of the loan that does.
As an example, suppose I obtain funding from the bank to pay for a new kitchen. I borrow that money in advance of the work being done (because I need to know that I can). But I don’t have to pay the installer until the work is completed. During the installation period, therefore, the funds sit in my deposit account. The loan has been made, but the associated reserve movement is several days or weeks later.
Now suppose that during the installation period, my wages are paid into the same account. When I pay my kitchen installer, is the money that I pay him from the loan or from my wages? You don’t know. There is absolutely no way of distinguishing between these two kinds of “money”. The drawdown of a loan is indistinguishable from any other sort of deposit withdrawal. People withdrawing funds from their deposit accounts – i.e. spending money – causes reserve movements: that effect is exactly the same whether the money in the account is lent to them by the bank or paid to them by their employer. So for a bank to get rid of excess reserves, it could lend, or it could encourage its depositors to withdraw (spend) the money that comes into their accounts from other sources. To a bank, they are the same thing.
To say that banks “lend reserves” is therefore looking at only one part of what banks do. Banks are deposit-takers and payment agents, too, and these functions ALSO cause reserve movements. There is WAAAY too much focus on lending and not nearly enough on what else banks do.
Not all loans are drawn, of course: refinancing loans, for example, where a bank is simply rolling over an existing loan, do not generate reserve movements. More problematically (for your argument), if a bank makes a loan which the customer then spends by transferring the funds electronically to another customer at the same bank, there is no movement of reserves. In this case, the bank has not in any sense “lent out” reserves. It has a new loan and a new deposit on its books, but its reserve position is exactly the same. Obviously the more concentrated the banking sector, the fewer reserve movements there will be: if there were only one bank, there would be no reserve movements at all, but that wouldn’t mean the bank wasn’t lending or people weren’t spending. Reserve movements in a highly concentrated banking system such as that in the UK (and that in Canada) tell us very little about the lending behaviour of banks or the spending behaviour of their customers.