Money, barter, the clearing house, and balance sheet recessions

I think David Beckworth is onto something important here. What looks like a balance sheet recession can in fact be caused by an excess demand for money.

The hairdresser cuts the hair of the manicurist, who does the nails of the masseuse, who massages the hairdresser. We have a Wicksellian triangle, with no double coincidence of wants. How can the three women coordinate their exchange of services?

1. Barter. All three women meet at once, and do a three-way barter deal to exchange services.

2. Money. Each pays money in exchange for services. Only two women meet at any one time. But if there is an excess demand for money, we get a monetary recession.

3. Credit. Each buys services on credit from the seller. Only two women meet at any one time. But the amount of credit grows over time. Eventually they reach a limit where they are unwilling to extend further credit, or unwilling to take on further liabilities, and we get a balance sheet recession.

4. Credit plus a clearing house. Each buys services on credit from the seller. But all three women periodically meet at once, to cancel out the circle of offsetting liabilities.

5. Credit plus money. Each buys services on credit from the seller. Periodically money circulates to pay off the circle of offsetting liabilities. Only two women meet at any one time. But if there is an excess demand for money, the circle of credits does not get paid off, and so eventually we get a balance sheet recession.

What lessons do we learn from this example?

First, we learn that credit is not a substitute for money. Only the combination of credit plus a clearing house is a substitute for money.

Second, we learn that credit plus a clearing house is like barter, in that it requires all three women to meet at once. Only money allows decentralised pairwise trades.

Third, as David Beckworth notes, what looks like a balance sheet recession can be caused by an excess demand for money. That same excess demand for money would cause a recession if there is no credit, or else cause a balance sheet recession if there is credit. We get a recession sooner or later in either case.

We should note one further thing. Money can facilitate trade even if there is no closed Wicksellian circle like in this example. Only money lets A buy from B, who buys from C, who buys from D, who buys from E,....in a never-ending chain. Barter, or credit plus a clearing house, cannot work if the chain never circles back to A, where it started. A might be dead by the time Z gets the money. The dollar never returns to the first person who spent it.

Comments

[primedprimate]

What causes fluctuations in the excess demand for money? Are fluctuations in the excess demand for money caused in practice by fluctuations in the supply of money or the fluctuations in the demand for money or both? Qualitatively, I believe the supply of money went up through the last recession but the demand went up even more. Nick, what do you think caused this most recent excess demand for money (or is that an irrelevant question if in the ideal world CB will always vary money supply to meet exogenous fluctuations in money demand)?

With the balance sheet explanation, there seems to be at least a superficial endogenization of the cause of recessions.

[Nick Rowe]

primed: it's a good question. But the only useful answer is: because central banks screwed up. Because they weren't targeting E(NGDP) ;-)

[HJC]

Nick: Isn't there a case in between 3 and 4 where a subset of the women can meet and trade credits? So, for example, the hairdresser can meet the masseuse and give over her outstanding credit from the manicurist to cancel her own debt.

[Tom Brown]

Nice post Nick, and thanks for directing our attention to David's as well. I saw his before... but skipped it, but it was well worth reading. David's stories are a bit racier than yours: you might have to sex yours up a bit to compete.

[Nick Edmonds]

1. What about transferable credit? E.g. hairdresser issues a voucher entitling the bearer to one haircut? Is that here?

2. The interesting question in case (5) is how much money is needed to clear the credits? Or whether you need any actual quantity of money at all, as opposed to just a monetary framework. Say all three women have outstanding debts of $100, they all bank at the same bank and they all submit instructions to settle their debts on the same day? Is any money involved? It's not like Beckworth's story where a something physical (transfer of a bill) has to happen in real time. Or maybe the bank here is just a kind of clearing house.

[Herbert]

1. Balance sheet recession is a red herring in my opinion: Government debt crowds out private debt, hence any "stimulus" is likely to result in smaller private liabilities.

2. The Great Recession may have been induced by an increase in money demand, or a fall in the cirular velocity. This is the most plausible explanation I am aware of.

[Nick Rowe]

HJC: "So, for example, the hairdresser can meet the masseuse and give over her outstanding credit from the manicurist to cancel her own debt."

and

Nick E: "1. What about transferable credit? E.g. hairdresser issues a voucher entitling the bearer to one haircut? Is that here?"

In both those cases, I would say that one of those women's IOU's was being used as a medium of exchange. It is money (at least within the circle of those three women).

Nick E: in your case 2, I would say the bank (or banking system) is acting as the clearing house. Money and the clearing house are much the same thing.

Herbert: I tend to agree. But it is interesting to consider the interaction between the excess demand for money and the balance sheet recession views.

[Phil Koop]

"What looks like a balance sheet recession can in fact be caused by an excess demand for money."

Sure it could. That would be the case if all those people with large mortgage and credit card debt had similar sized assets which in principle could be liquidated to pay off their debts. Are you arguing that lifetime labour is such an asset? Indentured servitude? If not, you might at least mention that this whole line of argument is a counterfactual in the context of the arguments that claim there exist balance sheet recessions.

"Money can facilitate trade even if there is no closed Wicksellian circle."

Steady on, Nick. There is no evidence that any physical quantity - time, space energy, the "multiverse", whatever - is infinite either in extent or divisibility. And while one is of course free to assert that a metaphysical quantity is infinite, it seems odd for an economist to place the economy in that category.

[BS Economist]

Forget Koo for a moment... I think that most of the people who talk about balance sheet recessions already have more or less this... I don't want to say model... idea in mind. I think the "balance sheet" part of balance sheet recession is just an explanation for how an excess money demand can become chronic.

As for why Koo doesn't believe that money expansions can't solve a balance sheet recession... I'm not sure I've ever understood that one.

[Nick Rowe]

Phil: consider two "balance sheet recession" models:

A symmetric model (like in my post) where everybody is in debt to everyone else, and hits a limit. Everyone stops spending.

An asymmetric model, where half the people are in debt to the other half, and it hits a limit. The debtor half stop spending, but what does the creditor half do? It's not obvious.

The real world will be a mix of both. It seems useful to explore the symmetric model. Everyone else (I think) has been looking at the asymmetric model.

We don't need time to be strictly infinite. Just that people are unsure in advance when the world will end, so money can't unravel by backwards induction from the final period.

[Nick Rowe]

BSE: " I think the "balance sheet" part of balance sheet recession is just an explanation for how an excess money demand can become chronic."

Hmmm. That's an interesting twist. I will mull that one over.

(I don't understand Koo either. It seems to me he really doesn't want monetary policy to work.)

[JKH]

http://macromarketmusings.blogspot.ca/2014/04/observational-equivalence.html?showComment=1398950165120#c215079548903336483

[Frank Restly]

Nick,

I get what you are saying regarding a balance sheet recession and the demand for money, however there also may not be enough willing sellers of goods (houses) for the given supply of money.

This is particularly true when an asset has a utilitarian purpose - in the case of housing, roof over one's head preferably close to one's work place. I may not want to sell my house at any price because I am unsure that I could find a similar house in the same general vicinity.

Balance sheet recessions occur when the value of an asset(for instance housing)falls relative to the value of a liability attached to that asset (for instance mortgage debt). That fall in the value of an asset can be caused by either people who are unwilling to part with money (excess demand for money) or people who are unwilling to part with the asset - even when a high price is offered.

This becomes more prevalent when the asset is subject to taxation. I may not want the value of my asset (house) to increase if I am being taxed on that valuation.

[PeterN]

"An asymmetric model, where half the people are in debt to the other half, and it hits a limit. The debtor half stop spending, but what does the creditor half do? It's not obvious."

Depending on the nature of the debt the creditor, forecloses, hires a debt collector, sells the debt to a debt collector at a discount and writes off the balance, sues to get a judgment and tries to collect on it, files to put the debtor in involuntary bankruptcy, files a bankruptcy claim if the debtor is already bankrupt or just writes off the debt as noncollectable (and I've probably missed a few possibilities - like default or workout of sovereign debt).

The common theme is that the debtor defaults, and the creditor has to write off part of the debt.

One person's loss is another person's forgiveness of debt. However this is an accounting asset, and is not always realizable. When it's not, the credit ceases to exist. Legal bankruptcy, foreclosure and insolvent estates are common examples.

Since effectively all of the reduction of consumer consumer debt has come from default, it's hard to argue that this has no economic significance.

And "one person's debt is another person's asset" an accounting identity from which it's dangerous to argue.

Perhaps you should distinguish illiquidity recessions from insolvency recessions.

[JKH]

I'd say a balance sheet recession in particular MUST be caused by an excess demand for money.

It's effectively an accounting tautology - as per my Beckworth comment above.

And more generally ANY recession MUST be caused by an excess demand for money.

Similar thinking.

And ANY recession is a balance sheet recession - IF credit exists.

It's only a question of degree.

What is being called a balance sheet recession lately is just an extreme case of a more general proposition - where credit exists.

Excess demand for money as a stock (M balances)

=

Deficient supply for money as a flow (V velocity)

[Frank Restly]

Nick,

On 5, do you know of anyone that has done a network stability analysis on interlocking rings of credit? Instead of one group of three women, you have a hundred or a thousand groups of three women. In a meritocracy, does the burden of debt shift to those with less merit? Can you correct the inadequacies of a credit based meritocracy with monetary policy alone?

I don't think you can. Presumably, those with higher merit will have the same access to monetary policy as those with less merit.

[Roger Sparks]

"I think David Beckworth is onto something important here. "

I would agree.

The story completes logically if money is considered as always being property. Each participant made a promise to deliver property in the future. When the property was finally available, it was delivered. The property was not consumed, and had no need to be consumed.

When we look at the story from a macro-economic standpoint, we discover a sequence of promises. Each promise is to exchange property for labor or material with the "money" part of the property transaction always uncompleted.

We are told in the story that there is a recession in the economy. There seems to be a rule that stops transactions at a limit of $100 of outstanding promises per participant but such a rule is not identified as the cause of the recession.

We are invited to choose between "a balance sheet recession" or "an excess demand for money." as causes of the recession in the story. The unannounced rule could underlay both choices.

Consider the choice of each participant at the moment of transaction. Each had something to trade and made a decision. Would they rather accept the promise of money or not make the trade? Each participant in this story preferred to accept the promise. Not only is money "property", a promise-of-money is also "property".

[Nick Edmonds]

Isn't Koo's point on monetary policy that efforts to increase the money supply will be hampered because it's normally lending that creates the money supply and in his recession, nobody wants to borrow. If you allow for monetary policy to include the central bank just buying stuff, maybe he'd see more scope for it.

You still need the money to get into the hands of the people with the debt of course, so it's likely to make a difference what you buy.

[jt26]

Hi Nick, is there another option?
A credit company that supplies short0term credit to cover time inconsistency at the clearing house. This credit could be backed by real assets as collateral for example? Does that mean we don't need "money", just real assets ... maybe this is the same as your "land as bubble money" meme. If so what does this imply for an excess demand for money?

[HJC]

Nick: "In both those cases, I would say that one of those women's IOU's was being used as a medium of exchange. It is money (at least within the circle of those three women)."
The original bilateral credit is also a medium of exchange (otherwise isn't it just barter?), so it's clear that it's also a medium of exchange if used in subsequent transactions. But is it money? Maybe it depends what units the credits are denominated in. If the credits are denominated in an unit from outside their circle then it can easily be imagined that the women may well discount the credits in their subsequent dealings depending on preferences or creditworthiness. Since one key feature of money is that it is always par in its own unit, the women's credits look less like money and more like a promise to pay money, even though they circulate like money (and bank deposits).

[Tom Brown]

Nick Edmonds,

You wrote: "Isn't Koo's point on monetary policy that efforts to increase the money supply will be hampered because it's normally lending that creates the money supply and in his recession, nobody wants to borrow."

I know very little about Koo. But is there anything wrong with viewing borrowing as just another market, and like every other market in a monetary economy, it's a market where non-money is traded for money? In this case the people supplying the non-money (and thus determining the supply curve for non-money) would be the borrowers, who supply units of loan principal in exchange for MoE (bank deposits). The lenders (the banks) would be creating the demand curve for those units of loan principal. The "units" are measured in the UoA. If viewed like that we have a normal supply and demand curve for those units of loan principal... and I'm sure if the terms of the loans were significantly beneficial to the borrowers they'd want to sell them. For example, people with underwater houses with high interest rate mortgages may in some cases have been happy to refinance them anyway, if they could get favorable terms. But the banks weren't buying loans on the terms the borrowers were capable of offering.

Although it does seem to me that there is at least one thing that can be different about the loan market from other markets: often the loan seller (the borrower) has the option of buying back arbitrary numbers of units of loan principal at par from the loan owner anytime the borrower wants. This is not always the case (prepayment penalties for example).

[Nick Rowe]

HJC: I think you are defining "money" as "unit of account". I was defining it as "medium of exchange".

[Min]

Maybe I am misunderstanding the idea, but I do not see how you can disentangle a balance sheet recession from a lack of money recession. I thought that in a balance sheet recession there is a tremendous overhang of private debt, so that people use more of their money than usual to pay off or pay down their debt. Well, paying down bank debt destroys money, so that is a money sink, a persistent drain of money out of the economy, which tends to perpetuate the lack of money. No?

[HJC]

Nick: Isn't that an unusual definition? Keynes made the distinction between 'money-proper' and 'convenient media of exchange' for example salt, beans, tobacco and so on.

[PeterN]

Min,

Interesting - Keen's debt accelerator in monetary terms. This has the big advantage of not violating the fundamental identity.

Of course not all debt is bank debt, and you have to consider default, but it seems an idea worth pursuing.

This post on debt dynamics is worth a look.

http://rortybomb.wordpress.com/2012/02/23/guest-post-by-jw-mason-the-dynamics-of-household-debt/

[Nick Rowe]

HJC: "Nick: Isn't that an unusual definition?"

Not unusual. But maybe a bit controversial. Even us market monetarists argue about whether it is the unit of account or the medium of exchange function that matters. My view is that for understanding recessions, it is the medium of exchange function that matters. But if the price of the medium of exchange were perfectly flexible, the real stock of medium of exchange would always adjust instantly so there would never be an excess demand for the medium of exchange, and no recessions. And the fact that money is usually unit of account also (or tied to the unit of account) makes its price imperfectly flexible.

Min: "Well, paying down bank debt destroys money, so that is a money sink, a persistent drain of money out of the economy, which tends to perpetuate the lack of money."

Good point. But it depends on how the central bank responds to that.

[Min]

Thanks, Peter, Nick. :)

[Nick Rowe]

Min: BTW: where did you get that point from? Did you read it somewhere, or did you figure it out for yourself, and figure everyone else must already have known it? (I know all the bits are out there, and commonly known, but putting them together, so paying down bank debt creates an excess demand for money and therefore a recession.)

[Nathanael]

This is an actual piece of economics work about the nature of money. One of the few I've seen.

It's amazing that it isn't one the first page of every economics textbook. Economics is basically about money, and this is a basic discussion of the function of money as a medium of exchange, and the effects of it.

Nick: I don't know where Min got it, but I also independently figured out the same thing. I'm not sure why this conclusion has been so *non-transparent* to most economists; I think perhaps because they're confused about the process of money creation and destruction.

Of course I also came to the conclusion that so-called "monetary policy" doesn't work to end recessions when there are a lot of poor people. This is because while "monetary" theoretically offers money, it only does so in the form of loans, so it doesn't supply wealth. People with no wealth are sensible enough to not borrow any more money, even though they *want* more money, because they recognize the crippling social effects of getting deeper into debt with no plan to pay it off. Even when the people with no wealth are willing to borrow, the banks aren't willing to lend to them. So it becomes impossible to increase the stock of money *among the people who need it* by lending -- it can only be done by giving them *wealth*.

This also seems bloody obvious and is also denied by lots of economists.

[Tom Brown]

Nathanael,

"People with no wealth are sensible enough to not borrow any more money, even though they *want* more money"

Speaking from the point of view of someone who's had a piece of property go "underwater" after 2008, I can tell you that I very much wanted to borrow more money precisely because I wanted to take advantage of lower rates to pay off an existing loan with higher rates... but I was precluded from doing that because the terms I was willing to offer the bank did not match the terms the bank was willing to offer me, so I could not sell a loan to them.

[Nathanael]

Tom Brown: yeah, thanks for the anecdote. That certainly happens.... and it is probably the more important factor...

"Even when the people with no wealth are willing to borrow, the banks aren't willing to lend to them."

I guess my point here is the old "pushing on a string" point. If you (as the central bank) are dependent on a set of commercial banks as intermediaries for your monetary policy, you can force them *not* to lend (by raising rates), but you can't force them *to* lend by lowering your rates. The commercial bank administrators can always decide not to.

The institutional structure is therefore critical. In the Great Depression, FDR bypassed this institutional structure problem by setting up quite a lot of banks which lent directly to people (and businesses), and were largely administered directly by the government.

[Prakash]

Hi Nick,

Slightly Offtopic.

It seems like the need for money is a positive function of the number of participants in the wicksellian circle, keeping communication, laws and trust constant, since every step will want to maintain a cash balance. Doesn't this imply that Nominal Gross output level targeting might be better than NGDPLT, since it will more closely reflect the demand for money for transactions?

[Nick Rowe]

Prakash: "Gross output" counts intermediate goods too, and not just final goods, right? Yes, you could make a case for doing that. But maybe we should count trade in used goods too? Dunno.

[Min]

Nick Rowe: "Min: BTW: where did you get that point from? Did you read it somewhere, or did you figure it out for yourself, and figure everyone else must already have known it? (I know all the bits are out there, and commonly known, but putting them together, so paying down bank debt creates an excess demand for money and therefore a recession.)"

Some years ago Brad DeLong put me onto John Stuart Mills's explanation of recessions as a lack of money instead of a general glut. So when I later heard about balance sheet recessions, that's what popped into my head. :)

[Nick Rowe]

Min. thanks.

[A.A.]

Time to reintroduce this?
http://www.globalresearch.ca/debt-cancellation-in-mesopotamia-and-egypt-from-3000-to-1000-bc/5303136

[Too Much Fed]

Nick's post said: Min: "Well, paying down bank debt destroys money, so that is a money sink, a persistent drain of money out of the economy, which tends to perpetuate the lack of money."

Good point. But it depends on how the central bank responds to that."

I don't believe it will be as simple as that.

[Jeff Cliff]

How can the three women coordinate their exchange of services?

6. Ripple (which potentially avoids the 'extra demand for money' problem pointed out above by primedprimate)

Frank Restly: 'Stability' is a common topic of discussion in the Ripple community. http://arxiv.org/abs/1007.0515 looks at one vantage point of what stability would look like (expressed in terms of failure)

Tom Brown: In the case of Ripple, money generation is an automatic aspect of how the system could work. It can arise in the case where people produce utility for eachother, consumer and producer surplus, in short any case where good outcomes happen and where there is no hostility. Even in the event of a tightening of supply and growing hostility, it is still going to happen to some extent, and that creates liquidity in times when it is needed.

A.A.: Unnecessary. Ripple shows us that there is a middle ground between jubilee and stable debt -- the choice on every level, by every person, whether a jubilee is called for is possible to resolve.

[Tom Brown]

Jeff Cliff, I'm not familiar with "Ripple" or the "Ripple community." This is what you mean?

http://en.wikipedia.org/wiki/Ripple_%28payment_protocol%29