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

"The supply (function) of money, and the demand for loans, together determine the quantity of money created, and that quantity created (eventually) determines the quantity of money demanded."

I agree with this, but feel that the demand for loans is itself partially a function of P and Y. At any given interest rate there will be a certain demand for loans. This will initially determine the qty of money. But if that qty of money causes changes in P and Y then this may in turn cause the qty of loans at that rate of interest to change, which will in turn change the qty of money supplied and so on until a new equilibrium is met (or perhaps until the money supply becomes extremely large or small?).

"Let us consolidate the commercial banks and central bank" ... As all right minded people do on occasion :D

Sadowski likes to give me a little bit of grief for that kind of thing, but overall I can't complain: he's generally game to answer my questions even though he grumbled about it a bit.

Nick,

"In this economy, the (change in) the stock of money is determined by the supply (function) of money (i.e. the rate of interest set by the central bank), and by the demand for loans."

In this economy, the change in the stock of money is determined by the supply (function) of money (interest rate charged by central bank for new loans), the demand for new loans, the amount of loans outstanding, and the interest rate on the existing loans outstanding. Remember in this economy no one ever defaults and there is one bank (the central bank). In this economy, all loans are repaid over time back to the central bank and so loan repayment reduces the stock of money flowing through the economy.

Also, under this situation it should be apparent that unless the interest rate on existing loans is a floating rate, the economy could suffer from an insufficient supply of money to service existing loans. Meaning, existing debtors are paying back money to the central bank at a faster rate than new loans are being created by the central bank. There is a potential for a "race to the exits".

"In this economy, the (change in) the stock of money is determined by the supply (function) of money (i.e. the rate of interest set by the central bank), and by the demand for loans."

Who determines the demand for loans?

If the bank is offloading loans as quickly as they make them , the money supply of interest is global. Interest rates will still have some effect , but the local CB generally loses traction.

"Spillover" is the hot topic these days , mainly for EMs , but advanced economies are not immune. EU hot money spilled over in the U.S. in the 2000s.

"The supply (function) of money, and the demand for loans, together determine the quantity of money created, and that quantity created (eventually) determines the quantity of money demanded."

Looks much like Godley and Lavoie to me, although they get there through a different exposition.

Nick,

You constructed a mini flow of funds model here, in which the change in the stock of money is coherently linked to the loan side from an accounting perspective. That seems very different for you (to me) and that's why it seems very post Keynesian in orientation. G&L basically assume away the concept of the stock demand for money as something that falls out of their models as determined by everything else. And that's why money has a supply determined quality (from a stock perspective).

Nick, I don't know if this belongs here or in the last post, but regarding Mark's complaints about the money multiplier bit in the BoE paper (he left a long comment in your last post), it seems he likes to point out this form of the money multiplier:

“Every Econ 102 textbook I’ve ever seen teaches the money multiplier is a function of three variables. The currency ratio is always portrayed as the depositors’ choice, the reserve ratio (above required, if any) is always the lenders’ choice, and the total amount of currency and reserves (the monetary base) is the central bank’s choice (even if supplied through the discount window), and all are dependent on the conduct of monetary policy by the central bank.” - Mark S.

So he means this:

1. c = currency ratio (borrowers’ choice)
2. r = reserve ratio (lenders’ choice)
3. MB = total of currency + reserves (central bank’s choice)

M = MB*(1+c)/(r+c)

Let’s analyze a simplified example. Say required reserves are 0% of deposits. And say cash is outlawed. Now our formula reduces to:

M = MB/r = (excess reserves)/((excess reserves)/deposits) = deposits.

So the CB chooses excess reserves and the lenders’ choose the ratio (excess reserves)/deposits. The borrower’s choice has been eliminated. But who chooses deposits?

Another way to say that is that lenders (who buy loans from borrowers) chose r so as to maximize their profits [Mark liked this sentence]. But isn’t the set of available r for them to chose determined by both the CB and the borrowers (the loan sellers) who control the supply curve for loans? [Mark's response to this sentence below]:

"Well, it’s kind of a simultaneous system with multiple markets each with a set of supply and demand curves and multiple agents each maximizing their utility. But in a very real sense the CB is running the whole show even if they can’t force the other agents to act in a certain way." - Mark S.

http://www.themoneyillusion.com/?p=26355#comment-323668

I *think* I like the 1st part of that about the "simultaneous system," but I'm not sure what it means? Any ideas? However the bit starting here "But in a very real sense the CB is running the whole show..." I told him I considered that to be an assertion... that at this point I'd have to take on faith because I don't see the full logic of it.

This business very much ties into your post here I think, and my question above? I think the answer to my question above is that the borrowers control demand for loans (i.e. the supply of loans to sell to the CB), and thus the do help determine the available choices for "r" which Mark says is the lenders' choice. What do you think?

TMF: "I agree with this, but feel that the demand for loans is itself partially a function of P and Y."

That would make sense to me. If we measure loans in nominal terms, the demand for loans should be proportional to P. It will very probably also depend on Y, though how exactly it would depend on Y would depend on what the loans are being used for, etc.

Tom: "Who determines the demand for loans?" My "model" isn't specified fully enough to say, other than "the people who want to borrow".

JKH: "Looks much like Godley and Lavoie to me, although they get there through a different exposition."

My immediate reaction was surprise. Then I did not find it surprising. Sort of ironic, but not surprising. Sometimes you end up in the same place via different routes.

But if I had assumed that the central bank buys apples, rather than IOU's, at a fixed price of apples, then I would have said that the (net) supply of apples, plus the supply (function) of money, determined the stock of money.

"But if I had assumed that the central bank buys apples, rather than IOU's, at a fixed price of apples, then I would have said that the (net) supply of apples, plus the supply (function) of money, determined the stock of money."

That's equally compatible with G&L (in my view).

The important thing I think is that it's a closed flow of funds model, and the stock demand for money is something that falls out as determined by everything else (my second comment above).

Apples are as good as gold.

JKH & Nick... am I reading you right? You're both agreeing that Nick and Godley & Lavoie are on the same page here? Does that tie into my question about the money multiplier above (March 16, 2014 at 06:17 PM)? My head is reeling here... ...Nick does this represent any change in your position? It doesn't sound like it from your last comment... so in some sense you've agreed about this all along?

steady, Tom - Nick is arguing from a starting point of considerable qualification:

"But I now want to argue for a more extreme position. I now want to advance deep into enemy territory. Suppose, just suppose, that the central bank does target an exogenously fixed rate of interest, and ignore the Wicksellian indeterminacy this creates, and ignore the fact that this is incompatible with targeting inflation or anything vaguely sensible."

Hi all

Sorry if this is a bit off topic, but it seems that in these discussions some key aspects of Post Keynesian thought are not being put forward so here are a couple.

Basil Moore (referenced in the BoE paper) was always quite clear that the CB rate, although he described it as 'exogenous', was never held static but was adjusted to meet the CB's policy objectives. Underpinning this, however, was the rejection of the Wicksellian idea of the natural rate of interest: the capital debates had shown that, in a heterogeneous-capital-good economy, it has no theoretical basis and is only valid in the one commodity (corn) economy with its production function parables.

So, ironically, 'exogenous' rates are perfectly consistent with a CB following a Taylor Rule (as one policy rule for instance). The endogenous vs exogenous labels were merely to provide verbal distinction from the various forms of Monetarism prevalent at the time. It may well be time for them to be abandoned since Woodford's models feature (what could easily be called) 'endogenous' money but, being neo-Wicksellian, are totally at odds with the Post Keynesian view.

Another interesting point to note is Kaldor's belief that there can be no excess of credit money - it simply does not come into existence! This was also supported under quite a lot of attack from fellow Post Keynesians (and Charles Goodhart) by Moore.

See the recent edition of the Review of Keynesian Economics for a mini-symposium on Basil Moore's book.

Nick,

On one of those linked posts, you said:

"The supply of money is determined by and equal to the demand for loans from the banking system at the announced rate of interest. The supply of money is not equal to the demand for money at the announced rate of interest."

So is the difference here that you're ignoring that "Wicksellian indeterminacy" for purposes of illustration? i.e. you're assuming temporarily that's not a problem?

JKH, duly noted, thanks!

JKH: "Apples are as good as gold."

Yep. But it is a bit trickier to apply it to a gold standard world, because sometimes the gold itself (and not just currency convertible into gold) can be used as a medium of exchange.

Tom: I confess I have not read Godley and Lavoie, so I cannot say. But I am familiar with some of Marc's work, and some of his students' work, so I have a general sense of it. It is not implausible.

This is no change in my position (well, not in the last few years anyway). I have several posts arguing sort of the same thing, in various ways. I stole the example of the perfectly interest-inelastic money demand function from my PhD supervisor, monetarist David Laidler (who is not to be held responsible for my misrepresenting that idea). Bill Woolsey (I think) will agree and find nothing new here (except maybe exposition). David Glasner (I think, but am less sure) will be less inclined to agree. Scott Sumner could go either way, and would ask why it matters. Steve Keen (I think) would agree, but would say it shows that money is endogenous, and would talk about the stock of debt created rather than the stock of money created. New Keynesians (I think) will disagree. Many Old Keynesians will disagree. Mike Sproul will strongly disagree. This stuff cuts across party lines.

It's a funny old world.

Nick,

In that same referenced post, you said:

"Perhaps we should abolish "the demand for money". It only confuses people."

Good idea.

E.g. it's not easy keeping thoughts straight about the stock demand for money when the flow of funds (e.g. bank loans creating money) can be a factor in satisfying an "excess demand for money".

O/T: JP Koning looks again at MOA and suggests credit card companies might have a competing one right now (especially there in Canada):
http://jpkoning.blogspot.com/2014/03/credit-cards-as-media-of-account.html

Fed Up or Too Much Fed, or whatever he goes by would be interested in that I think. :D

JKH @7.11. True. Godley and Lavoie would be (roughly) happy with those assumptions, I am not. I make them for illustration.

JKH @7:22. I don't think there's any substantive difference between what I am saying now and what I said then. Except I have cleaned up my language, and have just rapped my own knuckles for having said something so terribly sloppily.

"E.g. it's not easy keeping thoughts straight about the stock demand for money when the flow of funds (e.g. bank loans creating money) can be a factor in satisfying an "excess demand for money"."

Yep, that's why I did that old post where banks bought houses. Because it's harder to confuse houses with money.

HJC: I disagree on Cambridge Capital Controversies showing the natural rate of interest makes no sense (though it does have other problems, in a monetary economy with sticky prices). But let's leave that aside here.

But yes, there are definite differences, as well as similarities, between Neo Wicksellians and Post Keynesians.

The "exogenous/endogenous" distinction has become terribly unhelpful, because some people seem to use those words in very strange ways. Yep.

Thanks Nick, I think the main point I was trying to make (without having to actually discuss them) is that there were theoretical reasons for rejecting the problem of Wicksellian indeterminacy, see Colin Rogers' work for instance. And these were important parts of the monetary theory. So if there are any similarities between Neo-Wicksellians and Post Keynesians they are only very superficial. [Obviously, another related point is that it seems to me that a lot of this is not fully appreciated by many that put forward the 'endogenous' money theory, and this causes confusion.]

Nick, you write:

"The supply of money determines the quantity demanded, and not vice versa."

But does it solely determine the quantity demanded, or does it work in conjunction with the "demand for loans" which you said earlier (in your response to my question) was determined by "the people who want to borrow."

Nick, you write:

“Cutting the rate of interest would not work. But we know it will work, provided the quantity of loans demanded depends on the rate of interest.”

So the quantity of loans demanded depends on the rate of interest… but the demand curve which determines a quantity of loans demanded for each interest rate in turn depends on what? It still seems to me like you're that it depends on the “the people who want to borrow,” i.e. the people putting loans up for sale to the central bank, at least in part. Is that true?

Hi Tom, I think that what Nick means is that, first, the demand for loans determines the amount of money created - i.e. the supply of money. But although this supply is accepted by its recipients (because it's money), in all probability it will not match the desired money stock, or demand for money. So, second, P and/or Y will need to adjust to restore equilibrium on the demand side. Hence, by adjusting P and Y, "the supply of money determines the quantity demanded." All subject to stated qualifications. (That's what I think he means anyway.)

This question of the excess supply of credit money, that I referred to above, has been debated at length in the Post Keynesian journals.

JKH: "G&L basically assume away the concept of the stock demand for money as something that falls out of their models as determined by everything else."

That's where I would disagree. If people are holding more money than they wish to hold, they will spend more, to try to get rid of the excess. Collectively, they will fail to get rid of money, but their increased spending will increase P and/or Y. So we can't ignore the stock demand for money.

Tom: from my post: "The supply (function) of money, and the demand for loans, together determine the quantity of money created, and that quantity created (eventually) determines the quantity of money demanded."

In conjunction.

Nick, thanks I agree "In conjunction." So to answer my last question above:

Posted by: Tom Brown | March 16, 2014 at 10:41 PM

i.e. "Is that true?"

Sadowski & I were discussing it and I was trying to clarify:
http://www.themoneyillusion.com/?p=26355&cpage=3#comment-323881

The G&L models are based on buffer stocks which absorb the difference between each sector's expected and actual outcomes. Money is the buffer for consumers. (Inventory for firms etc.) At each time period the consumer adjusts their expectations, consumption and portfolio allocation plans until there is a steady state. At this point the ex ante supply and demand for money will be equated.

Yes you are right in your hypothetical example with the assumptions you make. But in the example you made you simply assume away the problem with the current system:

“Ignore risk on loans. Anyone who wants a loan from the central bank can get one, at the rate of interest r, and they never default.”

“By cutting the rate of interest, the central bank increases the quantity of loans from the central bank, which creates more money. Eventually P and/or Y will increase and the quantity of money demanded will increase in proportion to the quantity created.”

What happens if debt to gdp keeps increasing? Loan supply gets reduced because lenders are facing less quality borrowers. Borrowers want to borrow less also. Less loans equals less money.

The CB can keep bringing rates down further and further negative to compensate lenders. Debt to gdp keeps increasing and rates keep going further negative. At some point the monetary system collapses.

You need to take into account debt to gdp if we look at the current system.
In the long run bank lending based monetary policy fails because the banks misallocate funds during lending process so that debt grows faster than gdp.

Tom: in conjunction. True. (And stop using the words "supply of loans" when you mean "quantity of loans supplied"! The first is a function which describes what the supplier wishes to do, for various interest rates; the second is a number.) Mark S is right that PP and others like him will not like this post.

HJC: interesting. One old related post of mine. A second.

Your 10:45 comment was pretty accurate!

This works out in the information theory model:

Nick Rowe's model of the money stock

There's an interesting twist. Answering the question "Will it be true that the actual stock of money will always be equal to and determined by the quantity of money demanded at that target rate of interest?" ...

I will answer no, but add that the actual money stock will always be less than or equal to the quantity of money demanded. This is because you can only lose information that is sent in the original message (the demand).

Like I always say, "Print more money."

Especially now.

Jason, that is awesome! I don't know if you've got it right or wrong (I'm a rank amateur: and I just gave it a quick once over), but just that concept of what you've done is awesome... to take a paragraph from Nick's post and turn it into something more visual like that.

Now I'll read it and see if my non-economist mind can make any sense of it.

But I don't know how many times I've wished someone would do something like that.... hopefully you've got it all right!

Hey, while you're at it, what do you think you could do with this Nick Rowe comment: I'm still scratching my head over the "rectangular parabola" ... and pretty much all of it actually.

He's talking about a hypothetical I present here:
http://www.themoneyillusion.com/?p=26355&cpage=3#comment-323301
and I added "reserve requirements = 0%" too.
Nick's response was supply and demand for MOA both go to zero, and then he elaborated with the above. I'm still trying to figure out if that means that the new steady state price level will thus not change (i.e. not go to zero) or not.

Man I wish I could visualize what Nick was talking about there with the rectangular parabola... I was going to take a crack and drawing it myself and see if I could get him to take a look.

Thanks Nick. A brief comment on each link:

I think that Keen's model is a strange twist on Moore's observation that for aggregate demand to increase there must be 'deficit spending'. Credit money (as well as previously-saved money balances) can be the source of this spending. But, since credit money is 'endogenous', it does not cause the spending increase (or anything else!). Keen seems to have missed the second point and his ideas are really some form of monetarism.

Personally, I like Perry Mehrling's idea that monetary policy affects asset prices in the first instance - any effect on borrowing, spending, investment etc come (perhaps much) later. This also seems to me to form a better basis for a 'one general theory of money.' (Add in a suspicion of the validity of the Wicksellian concepts as well.)

Tom: it's rectangular hyperbola: Y.X = some constant.

HJC: "Keen seems to have missed the second point and his ideas are really some form of monetarism."

Maybe that's why it sorta resonated with me!

Nick, thanks!! Whew! ... I'm pretty sure Mark was just messing with me at the end there... the smiley face says he was. I'm happy to take your word about Philip... and Mark's, the thing is I don't know anything about Philip. People must assume I'm a fan of Philip.

And regarding this: "And stop using the words "supply of loans" when you mean "quantity of loans supplied"! The first is a function which describes what the supplier wishes to do, for various interest rates; the second is a number."

I absolutely agree!... I thought I was being careful, but I guess not (I'll have to re-read and see where I misused it). I know "supply" is a function, but every point on that function (the locus of points) is determined by... What? "the people who want to borrow" ... and some other stuff? Am I still missing something here. Shoot! :(

Nick: Yes, but you're not a soi-disant Post Keynesian! I was hoping to get your opinion on Perry Mehrling's work, perhaps a post in the future? (You did a great job of discussing John Cochrane's papers after I mentioned it, thanks.)

Nick & HJC,

Mark A. Sadowski already proved that Steve Keen was not only a monetarist, but an OLD school monetarist here:
http://www.themoneyillusion.com/?p=26140#comment-318544
:D

"Tom: it's rectangular hyperbola: Y.X = some constant."
Ah!.. I thought I was OK with my quadratic surfaces, but I must have missed the day they named that... I thought it was something you made up! ... or an econ thing. Ha!
http://en.wikipedia.org/wiki/File:Rectangular_hyperbola.svg

Tom: Oh dear, if it weren't so sad, it would be funny. Thanks for the link.

Tom: Sorry, the comment above is about Sadowski's observations on Keen and not your re-discovery of the rectangular hyperbola.

HJC... Ha!... no problem. You wouldn't have been the 1st to give me a hard time today (not that you did at all).

BTW, I go into some detail trying to figure out what Keen was talking about. I do it with Sadowski, but I have no idea where that thread is now. But I have the gist of it here:

http://pragcap.com/forums/topic/steve-keen-an-old-school-monetarist/page/2/#post-61368

Starting with "Oilfield, thanks. There’s a few issues there."

"Oilfield Trash" was generally defending Keen, in particular from the charge of being a monetarist (which I never took seriously anyway), but I still couldn't figure out what Keen was trying to do in those few formulas. Any idea?

Cheers, Tom. Thanks. FWIW, I don't know if I am right either. I saw the later comment and you seem to be in the right place on the rectangular hyperbola ... the solution Nick was talking about is the highest point (1/P = infinity) on the demand curve (rectangular hyperbola) that occurs as you approach zero one the x-axis (quantity supplied).

Jason, thanks so much for taking a look. It's starting to make a little more sense now, but I'm not there yet. So the x-axis has quantity (supplied and demanded) on it. True?

Now I know "price" is normally on the y-axis, but in this case, I'm not sure what that means exactly. Nick writes:

"The vertical axis has (say) 1/CPI."

I know what the CPI is... but what is actually on the y-axis here? If we're at the tick mark labeled "10" on the y-axis what does that mean? \$10? 1/(\$10)? 1/(\$0.1)? I'll guess 1/(\$0.1). And I guess Nick means it's normalized by CPI instead. So 1/(0.1*CPI) then? or CPI/(\$0.1)?

OK, some background: Scott thought supply went to zero (only), and Nick thought demand went to zero (originally), but then later Nick said both supply & demand go to zero as my epsilon goes to zero. So if both go to zero, then Scott's solution (P goes to 0) is not correct. So what is the solution then? Is the demand curve a rectangular parabola now still? And if it is, then what does "making it go to zero" look like? Does it mean the rectangular parabola more closely hugs the x & y axes? (i.e. y = K/x, as K goes to 0)? Does it mean K stays the same, but it just shifts right or left or up or down? I'm going with K goes to 0.

I think I have the supply curve figured out: it's a vertical line, e.g. x = Qs. To make it go to zero means to move it left towards Qs = 0.

And regardless of how the curves look, does it mean the price level stays fixed as supply and demand (i.e. "epsilon" in my example) go to zero?

Thanks!

Hi Tom

I've had a quick look at the long quote posted by Oilfield Trash. My impression is that Keen is confusing the separate concepts of income/expenditure and cash-flow/payments. Expenditure equals income at the aggregate level but the parallel system of financing this is quite separate since payments can be delayed by credit and all sorts of financing arrangements. His argument that debt injections create a discontinuity in expenditure/income suffers from not keeping the profit and loss accounting separate from the cash-flow accounting.

I'm not sure, but he also seems to confuse 'Loanable Funds' with a denial that banks create money. In his model banks don't seem to have pay interest on the deposits they create either.

HJC, thanks for checking that out. I'm not familiar with the income/expenditure or cash-flow/payments you describe. Oilfield brought that up? I'll take another look.

My problems with Keen's paper had to do with why he talks about an instantaneous AD(t) = V(t)*M(t), then integrates it (incorrectly) over a period of time. And then at one one point multiplies by a time interval, etc. I couldn't figure out if he was trying to calculate a cumulative AD or an average AD, or even if those concepts make sense. Also he seems to be saying AD = Y instead of Y*P. Why? I'll re-read and see if I can find the part about the points you were making... plus I didn't read all of Oilfield's refs yet.

Tom: I'd need to look again and focus on the instantaneous aspects. But I think the problems are more fundamental than whether the maths is correctly done or not.

"the interest rate target is endogenous with respect to the stock of money, if inflation is to be kept on target"

The interest rate is not endogeous, it is exogenously set by the central bank according to their theories.

If you had terminator robots patrolling the streets programmed to shoot people whenever CPI inflation rose, would that make the shooting of people by terminator robots "endogenously determined"? Nope.

Philippe: central banks adjust the money base to hit their interest rate target, and adjust the interest rate target to hit their inflation target. Everything they do is "according to their theories", and depends on what is happening in the economy.

Is the weather endogenous? In an economist's model, the weather is (usually) exogenous. In a meteorologist's model, the weather is endogenous.

"If you had terminator robots patrolling the streets programmed to shoot people whenever CPI inflation rose, would that make the shooting of people by terminator robots "endogenously determined"?"

Yep.

Steve Keen and hot potatoes:

Start in equilibrium. Then the bank cuts the interest rate. The representative agent plans to borrow \$100 more and spend \$100 more than his income. But he does not know that he is the representative agent. (Why should he, unless he knows everything about everyone?). So when he spends an extra \$100 he is surprised to find his income is also \$100 higher. He was "fooled" into borrowing that extra \$100. He could have spent an extra \$100 without borrowing anything. He was "fooled" into holding \$100 more money than he planned to hold. He revises his expectations and his plans. But he does not know if that extra \$100 income was a permanent increase, just a temporary blip, or was just some of his customers buying stuff a few days earlier than normal.

Modelling this formally, with proper math, would be hard. You would need aggregate and individual shocks, with the agents unable to distinguish the two. You would need temporary and permanent shocks, with the agents unable to distinguish the two. You would need the central bank to sometimes make a mistake in cutting the rate of interest, because the central bank has imperfect information too.

Standard Keynesian and New Keynesian models ignore all this. They just assume the representative agent knows he is the representative agent, and so knows his own plans and knows his own income.

I'm not surprised Steve Keen is struggling with the math.

Probably the simplest way to set up the math is like this: Discrete time model. Within each period, this is the order of play:
Agent borrows from the bank. Agent spends money. Agent learns his income. In other words, there is a short delay between the time everyone spends and the time everyone learns how much everyone else has spent on his goods. They don't add up the sales receipts until the end of the period. I reckon that's sorta what Steve Keen is trying to do.

Nick, re: Keen & math: Yes, ... and some people are just now realizing what a rectangular parabola is (can you believe it?). BTW, that post you did on Keen was identified by a commenter on Glasner's latest, pointing out that you, "of all people," get why Keen needs to resort to Lebesgue integration. Which led me to this very funny post:
http://fieldsfinance.blogspot.ca/2012/10/of-course-its-model-duh-final-post-on.html

Nick,

"Yep".

Nope. Rising CPI inflation does not cause terminator robots to shoot people in the street. Terminator robots shoot people in the street because evil robot lords program them to shoot people in the street. If the robot lords decided not to send out their terminators no terminators would go on murderous rampages. The robot lords decide to send out their terminators because they are crazy and evil and think that this is the right way to control inflation.

Rising CPI inflation does not cause the central bank interest rate to rise. The central bank interest rate rises because the central bank decides to raise it. If the central bank decided not to raise the interest rate the interest rate would not rise. The central bank decides to raise the interest rate because they think this is the right way to control inflation.

Any similarity between evil robot lords and central bankers is purely coincidental.

Philippe, what if the evil robot lords all die off, but no body can figure out how to turn off their self-sufficient terminator robots, and they continue shooting people in the streets every time CPI rises from now till the end of time? Still "Nope?" And what if additionally it turns out the so-called evil robot lords were actually just a group of monkeys fooling around in the lab, who just accidentally programmed the terminator robots? Still "nope?"

I was interested to read that new Keen paper. It clearly inspired by Nick's earlier post (that he reprises above).

Keen now thinks that AD increases by the change in debt * V. Isn't the "hot potato effect" implicit in his V ? I borrow \$100. The \$100 then "hot potatoes" around until we reach a new equilibrium. That's just velocity of money in action applied to an increase in M , right ?

The odd thing is that this V presumably includes people parking some of their new higher income in the bank where it will be lent out again and further add to AD. There seems to be an implicit money multiplier hidden in Keens new model.

Its also odd that Keen ignores govt deficit as a source of increase in AD.

The Market Fiscalist, Mark Sadowski pointed out that Keen is implicitly assuming V is constant... or that he's doing the math wrong:
http://www.themoneyillusion.com/?p=26219#comment-319483
Holding V constant is part of why he joked (see my above link) that Keen is now an old school monetarist.

Philippe: "Rising CPI inflation does not cause the central bank interest rate to rise. The central bank interest rate rises because the central bank decides to raise it."

By that same argument, a higher price of apples does not cause me to buy fewer apples. I buy fewer apples because I decide to buy fewer apples. The quantity of apples I buy is exogenous.

Tom: don't tell anyone, but I don't actually know what Lebesque integration is. (Or, I don't think I do). I'm pretty sure Steve Keen held V constant for simplicity, just like I did here, because it didn't really affect his point much.

TMF: "Keen now thinks that AD increases by the change in debt * V. Isn't the "hot potato effect" implicit in his V ? I borrow \$100. The \$100 then "hot potatoes" around until we reach a new equilibrium. That's just velocity of money in action applied to an increase in M , right ?"

Yes. I think that's right.

"Its also odd that Keen ignores govt deficit as a source of increase in AD."

Presumably for simplicity, because it didn't affect his main point, just like I ignored it here.

Nick, your Lebesgue secret is safe w/ me! (BTW, I suspected that might be the case, but I didn't want to presume and put a smiley just in case). Re: V constant: maybe you're right, but given that he spends effort making the rest more complicated, it's hard to believe... and not only that but Sadowski is right!... hold V constant and Keeen's math reduces to the exchange equation w/ constant V. Old monetarist, no?

> That's where I would disagree. If people are holding more money than they wish to hold, they will spend more, to try to get rid of the excess. Collectively, they will fail to get rid of money, but their increased spending will increase P and/or Y. So we can't ignore the stock demand for money.

Is that necessarily true?

In this model, people can't hold "more money than they wish to hold." They can hold more loans, however, which carry an expectation of repayment. That restricts the ways an agent can individually get rid of money, since spending on consumption does not offer the ability to repay the loan. Eliminating default risk doesn't really help, because to make the model well-defined we also have to assume that agents don't behave in ways that would inevitably lead to default.

This suggests that even if loans are "forcibly" given, the resulting money doesn't necessarily have to enter circulation. Although it may exist on a ledger somewhere, it's part of the money supply only in the same way that a lost and buried trove of gold coins is part of the money supply.

The flip side is that such a monetary policy can have *distorting* effects on the price level -- if forcibly given a 0% loan I will seek some sort of less-liquid savings vehicle that offers a >0% return. That still won't be consumption, but it does mean that I would be tempted to buy property or capital stock. In equilibrium, 0% loans would drive the return on all investments to a 0% risk-adjusted rate, at least over the same durations where bank loans are available. (A 30-year instrument would still have >0% yield if 0% CB loans were only available over 30-day terms.)

If such investment opportunities are not available for agents (that is, we're at equilibrium), then the excess of loans will result in an excess of sat-upon cash. With banks (in the US, anyway) holding on to much larger-than-historical reserve levels, there's some indication that we're here.

The central bank may be able to enforce a quantity of money, but in so doing the stock of money loses its meaning for the rest of economics -- its actions here cannot unilaterally affect the stock of *circulating* money.

"Suppose, just suppose, that the demand for money were perfectly interest inelastic. Desired velocity is fixed. So the money demand function is Md=L(P,Y)=PY."

But what if Md = PY/V? And suppose V changes. I'm thinking of a basic IS-LM model with LM given by M = PY/V. If you work out the equation that shows how M depends on the model's exogenous variables under a fixed interest rate policy this equation will have V on its right-hand side (as well as everything else that shifts the positions of the IS and LM curves). But then how can we say M is independent of the demand for money?

The case we're considering here reminds me of the IS-LM analysis of an economy under a fixed exchange rate. In that set-up Y is determined by only the IS curve and the world interest rate r*. The demand for money doesn't enter into it. That's because under a fixed exchange rate the central bank must automatically accommodate any change in the demand for money by changing the quantity of money. Here again, Y doesn't depend on the demand for money but M does.

Most of the above discussion went over my head, so apologies in advance if this has already been deal with.

HJC, "But I think the problems are more fundamental than whether the maths is correctly done or not." I agree, I'm wondering what the point of some of the math is... it would make it easier to evaluate its correctness.

This current post has made me realize the commonality there is between the "endogenous money" theorists and the kind of monetary theory done by Nick.

Both recognize that an increase in the demand for credit can drive an increase in M , AD and Y. Nick focuses on the generic affects the increase in the money supply will have and the control mechanisms that can be used to maintain stability, while the "endogenous" guys like Keen tend to focus on the increase in credit as something worthy of independent study (I think they follow Minsky in seeing debt-levels as one of the main drivers in the business cycle ?).

Question: is there any fundamental difference between a change in AD due to a change in V, compared to change in AD due to change in the demand for credit? I suspect Keen would say there is because (he would argue) in the run up to 2008 there was an unsustainable increase in the demand for credit that must (logically) have been accompanied by a decline in V since AD stayed on a steady growth path for the 2 decades before.

O/T: which English speaking regions use "maths" ... I'm from California, and I swear I'd never heard "maths" (only "math") until the last few years. Have I been guilty of a regionalism this whole time?

Nick, I checked where I used "supply" and I intended the curve every time. Did the curve not apply in some cases?

Tom,

British use 'maths'

Australians use 'Lebesgue Integration'

The Market Fiscalist,

"Is there any fundamental difference between a change in AD due to a change in V, compared to change in AD due to change in the demand for credit?"

Careful here. I think what you mean to ask is "Is there a fundamental difference between a change in AD due to a change in liquidity preference, compared to a change in AD due to a change in demand for credit?"

The liquidity preference of borrowers is presumed to be zero - they borrow money to immediately spend it. The liquidity preference of holders of money can be anything from zero to one.

And so AD and velocity are affected by both credit demand and liquidity preference.

JKH, Funny! ... BTW, Matheus's post was funny too... he sounded like he'd been harried by the "accounting police." Did he ever win them over?

JKH,

Lol. Indians use Dirac Delta Functions :-)

Tom,

Yes Brits use 'maths'. For you it is ok but India has got Americanized and I feel inflamed if Indians used 'math' because in school here we say 'maths' (although I have made this mistake too). Worst was an Indian school teacher Twitter status I saw which said it was always math and not maths!

The Market Fiscalist,

"This current post has made me realize the commonality there is between the "endogenous money" theorists and the kind of monetary theory done by Nick."

Allow me to quote Mark Sadowski:

"So, in short, there is no such thing as “exogenous money” theory. (Where's the Wikipedia page?)"
http://tinyurl.com/ot7k9s7

Majromax: each individual borrows money from the bank, planning to get rid of it, by buying something from another individual. But he doesn't realise that every other individual is planning on doing the same thing. So at the end of the period he is surprised to discover that for every extra \$1 he spent he got an extra \$1 in sales (on average across individuals), so his stock of money is higher than he had wanted it to be. Individually each one borrowed what he wanted to borrow. Collectively they all borrowed more than they wanted to borrow, and now hold more money than they want to hold. So the process doesn't stop there. But how exactly the process unfolds from then on depends on how they revise their expectations.

Only if individuals were indifferent to how much money they held, regardless of income, spending, interest rates, anything, so their current stock of money had no effect on their plans, could we ignore the demand for money and their actual stock of money.

Maurice: Start in equilibrium, then suppose V decreases (but nothing else changes). I think (given my assumed money supply function) the only effect would be an increase in loans from the central bank, where people borrow the extra money they want to hold. In this case only, the increase in quantity of money demanded *would* cause the increase in the actual stock. They borrow extra money because they want to *hold* extra money, rather than *spend* extra money.

When I said: " When people borrow money, they (usually) borrow it to spend it; they do not (usually) borrow it to leave it in their pockets." This is the "unusual" case.
I think you are getting it.

The ISLM representation is a bit weird. We have a vertical LM curve for a given M, plus a horizontal LM curve for a given r, and the vertical LM curve will shift right over time if the bank is making positive net loans. But the economy will be "off" both the IS curve and the (vertical) LM curve in the short run, because actual income does not equal expected income, and actual M does not equal M demanded.

TMF: Yep. I focus on M. "Credit" matters insofar as it influences M. Individuals borrowing and lending amongst themselves doesn't matter (much). It (mostly) only causes a redistribution of demand.

On your question: see my response to Maurice above.

Ramanan, I'm used to being taught how to speak my language by foreigners. Californians in particular could use a bit more of that. Worst was a student from Germany lecturing me on the proper use of "who," "whom," "could," and "would." He was right of course.

"Philippe: central banks adjust the money base to hit their interest rate target, and adjust the interest rate target to hit their inflation target. Everything they do is "according to their theories", and depends on what is happening in the economy."

Nick - I'm on look-out here :). I thought we had reached an agreement that hitting a central bank's interest rate target is not just about adjusting the monetary base. It depends on the system and tools the central bank choose to employ. It should be readily acknowledged that countries that operate a symmetric channel system try not to adjust the money base to move the target. The whole point of the system is to make the central bank's life easier by allowing them to mostly just change their administered interest rates to hit their target - this was all made clear in that Woodford paper, among many other articles.

(Of course acknowledging the adjustment of quantity to stabilize a given rate.)

Nick, in your example of 10:12, isn't the final stage that the RA pays back the \$100 loan with the cash-flow from the extra income. Even in an equilibrium case there is a need for this kind of circular flow of credit because of the mismatch in timings of expenditure and income. Surely that's what credit money is for. But how is it a hot potato?

On another matter, a key aspect of the pure-Endogenous money theory is the reversal of the causality in the equation of exchange. Clearly Tom's comment above has shown that this is not the case in Keen's model - it is more accurately described as an (old school) monetarist model. From the early days of the theory it was always claimed that money was an effect not a cause.

I think Keen is making a mistake (category error?) in trying to add new financial debt, which is a money/credit cash-flow, to expenditure which is a contract for exchange, i.e. another form of credit. They are in different realms, the former is used to settle (with respect to the seller anyway) the latter. It's a form of double-counting where basic concepts of income statements (or profit and loss accounts) are confused with cash-flow statements.

Nick, I think I've got my epsilon problem. Real demand for money = Mdr, nominal demand for money = Mdn, nominal supply = Ms, and the price level starts off at P0 = CPI. On the x-axis (dependent variable) is nominal demand and supply for money, while y-axis (independent variable) is 1/P. What's plotted is Mdn as a function of 1/P with Mdr fixed: Mdn = Mdr/(1/P). The supply curve is x = Ms = epsilon*Ms0, i.e. a vertical line at x = Ms. As epsilon approaches 0, real demand for money also falls proportionately as Mdr = epsilon*Mdr0. Starting out at epsilon = 1, we know that Ms0 = Mdn0 = Mdr0/(1/CPI), which is still true as epsilon goes to zero because epsilon*Ms0 = epsilon*Mr0*CPI. Thus P = CPI still (P does not change). What are some examples of minimal changes I could make to my hypothetical so that Mdr remains fixed at Mdr0 instead of being proportional to epsilon? Introduce a second commercial bank? Something else?

HJC, looking at it again, Mark did make some simplifying assumptions (he was clearly making a joke more than anything else). But he did accurately identify that deltaV = 0 (in the 2nd link to Mark I provided), else Keen's math wouldn't work. But Nick could be right... Keen does keep V a function of time, so perhaps he assumes it changes very slowly, ignores the deltaV terms as small, and just adjusts V to to the new time interval?

Hi Tom, it doesn't really matter too much about what his assumptions are regarding V. If he assumes that the causation runs from MV to PQ/Y then it's not really a Post Keynesian model, it's more of a monetarist model.

Out of interest, how many Tom Browns are there commenting on this blog? Or is there just one that never sleeps?

Just one insomniac. :D

ATR: there are many many ways a central bank could target inflation. For example, it could buy and sell gold, and adjust the target price of gold up and down, like Irving Fisher's "Compensated Dollar" plan. If Philippe is going to argue 'but real world central banks adjust an interest rate target' to keep inflation on target, then I will adopt the same mode of argument and say that the real world Bank of Canada (nearly always) keeps those spreads fixed at 25 basis points! It *could* vary those spreads, as you suggest, but it could also buy and sell gold, or farmland, or whatever.

And if they used a varying land price target to hit their CPI inflation target, they would never have to worry about the ZLB, where they have to abandon interest rate control and switch to "QE".

HJC: "Nick, in your example of 10:12, isn't the final stage that the RA pays back the \$100 loan with the cash-flow from the extra income."

That depends on a lot of things. If there is still a positive net (flow) demand for loans from the central bank, because the rate of interest is lower than equilibrium, that won't happen. To restore (ISLM) equilibrium, we need the representative agent to have planned expenditure equal to his *expected* income. Every time he changes his plans in response to surprises in his actual income, he changes his actual income. It is not at all obvious whether and how this learning process will converge to an equilibrium.

There are 2 LM curves: one for the interest rate set by the central bank, and one for the existing stock of money. What process ensures they both cross the IS curve in the same place? Especially since it's not at all obvious the IS curve slopes down, given the accelerator effect on investment of increasing income. The central bank may need to deliberately stop the process, by raising the rate of interest.

Sure - fine. You just sounded particularly attached to the need to change some sort of quantity there. Although I'm not quite following why you think keeping a fixed spread but moving it around (e.g., 1-2% to 3-4%)is a counter to the idea of changing the width of the spread. Moving a fixed spread can still move the interbank rate to a new target without adjusting quantity of reserves, all else equal. Or maybe I'm just missing your point...

Nick: For me to be able to understand this, I might need to resolve some of these issues: Who is the RA borrowing from? Why would there be a positive flow of loans when all agents are the same? I can see that we are not using rational expectations, but what about some sort of adaptive expectations where the agent eventually realises that his expenditure and income are always the same? Can the agent buy assets from the bank/lender to get rid of unwanted deposits to restore equilibrium via the asset markets? As it stands it's hard (for me) to appreciate what insights can come from analysing a model that wasn't really designed to have a natural place for inside money. But I think this may be all too much for a comment chain!

The idea of two LM curves is interesting, thanks. I've managed to wander quite far from simply correcting a few Post Keynesian representations.

HJC: With identical individuals, the representative agent would know that if he was planning to increase spending by \$100, then he would know every other agent was planning to do the same, so would expect his income to rise by \$100. But if there are aggregate shocks and individual-specific shocks to planned expenditure, and each individual only observes the sum of the two shocks, the representative agent would not know he is the representative agent. He gets a shock and thinks it is (partly) specific to him, so when he increases his planned expenditure he does not expect his income to rise by the same amount.

With imperfect information, and enough different shocks (aggregate/individual, and permanent/temporary) this is all compatible with rational expectations.

My mental model has each agent being able to repay a loan at the bank any time he wants. But if the interest rate is low, he won't want to. Instead he spends his excess money. And then is surprised to find it comes straight back to him, in the form of higher income! Income rises until the extra stock of money is willingly held.

Nick: This is what I think you are saying: the agent, because the interest rate is low, just wants to keep on spending, never paying back the original loan, even though his income has increased. But since the other agents are doing this too he also has an unexpected increase in deposit balances. Everyone's income has increased. But instead of repaying the loan with his increased balances, he spends it (again).

But the rate of borrowing and spending is not ever-increasing, is it? It's just a one-off increase, because the agent doesn't know whether the increased income is permanent, even though it is. My question is: isn't this just the same as saying that the agent receives an unexpected \$100, repays his loan, then, because interest rates are *still* low, borrows \$100 to spend, repay etc. When put this way is there really anything that can be called an increase in the money stock?

HJC: Suppose the central bank bought and sold apples, at a fixed price of apples, instead of buying non-monetary IOUs, at a fixed rate of interest. There could never be an excess supply or excess demand for apples, at that fixed price of apples. Anyone who held too many apples or too few apples would immediately sell them to the central bank or buy more from the central bank. But does that mean there cannot be an excess supply or excess demand for money? There could still be an excess demand or supply of all other goods, if the central bank set the price of apples too high or too low (assuming other prices are sticky).

For example, start with all markets in equilibrium, then suppose the central bank lowers the target price of apples. People buy apples from the central bank, and the stock of money shrinks, and (unless the prices of all other goods fall immediately in proportion) we get an excess supply of all other goods, and an excess demand for money, in all other markets. Setting the price of apples too high causes an economy-wide recession because it reduces the stock of money in circulation.

We find it hard to distinguish between the demand for loans from the bank (the supply of non-monetary IOUs to the bank) from the demand for money. It is much easier to distinguish between the supply of apples and the demand for money. That is why this old post I wrote makes the same point more clearly than this one does, even if it is less "realistic".

Nick: I can see that I'm going to have to spend some time thinking about apples. Will return when (if!) I have anything useful to say about them. Many thanks for your patience and time.

so i think the idea of money supply can be thought of as the money out there as cash, or at least very liquid and this is important in monetary theory, because that cash will be looking for a place to get return, interest rate versus marginal efficiency of capital as per keynes

but money supply also should be thought of as all the total money that could be liquidated,

bonds owed by the government that could be cashed in, loans backed by government, etc

(I dont think it should include the IOU money that banks create by, you know, not actually transferring the money, so the same money supply can be loaned several times, supposedly increasing the money supply.... because both the interest and principle paid on the iou money loans must be in real money (backed by government))

because where as monetary stimulus is basically increasing the liquid money supply

fiscal stimulus is a transfer of percentage of money supply from low multiplier place to high multiplier place, so for that purpose you cant just consider cast in circulation, but total money supply

I think thats basically what keynes was getting at

Test

Ms=8.33 @ epsilon=1

Nick: I'm on-side with your apples' example. One thing I would be interested to know is how much of the adjustment to restore equilibrium demand for money would happen in the capital markets rather than the goods markets. Your example above relies somewhat on price stickiness, which is more a feature of goods markets. Capital market prices move continuously and can adjust readily to maintain any arbitrage/equilibrium relationships.

I wonder what you think of Ramanan's post on this.

HJC: Thanks. On reflection, I think my apples story was the clearest.

On price stickiness. Suppose all prices were sticky, except the price of peanuts was perfectly flexible. So the peanut market always clears. We would then blame recessions on the price of peanuts being too low. The peanut theory of recessions. For "peanuts" read "bonds and other financial assets", which do indeed tend to have very flexible prices (most of them), probably because they are homogenous with lots of buyers and seller.

Not sure where to find Ramanan's post.

Nick, I just clicked on his name above (Ramanan's):
http://www.concertedaction.com/
The top post looks interesting, but I don't know if that's the one

Tom: Yes , that's it: 'Reconciliation Of The Supply And Demand For Money'.

Thanks Nick: Now I'll think about peanuts for a while...

HJC and Tom: thanks for that link. Yep. Very interesting, laying out the different answers to the exact same question I'm bashing my head against. Like I said, this stuff cuts across the usual "party lines".

My underlying position is that money is not like other assets. It doesn't have a price of its own, it does not have a market of its own, and the demand for money is a demand for a buffer stock inventory that everyone both buys and sells whenever anything else is sold or bought. So when there is an excess demand for money, trade in all goods (not just newly-produced goods) slows down. And we can't just look at any one market to see if there is an excess demand or supply of money. Money is weird, and so the answer to the question: "what (if anything) ensures Md=Ms?" will not be like the answer to the same question for any other asset.

For example: take a standard Keynesian unemployment "equilibrium", where all markets are clearing except the labour market. Is there an excess demand for money? Well, yes, because unemployed workers want to buy money in exchange for labour, and cannot. But that does not mean they want to "hold" (all) that extra money, except temporarily. They want to spend (most of) it.

Nick: Did you also look at the link to Peter Howells' paper? It's a good summary of the internal debates within the Post Keynesian tradition, you might find you have more in common than you realise.

I'm with you on your underlying position, and I have recently been thinking a lot about how Perry Mehrling looks at this stuff, which is why I keep mentioning him. At this stage I don't understand why he doesn't get more attention.

HJC: I have just now read the Howell paper. It is very good. Two points:

1. Minor point, but he should have his knuckles thoroughly rapped for confusing: supply (curve); quantity supplied (quantity offered for sale, that sellers would like to sell); quantity sold. His example of haircuts is totally wrong. Just because the quantity of haircuts sold equals the quantity demanded does NOT mean the haircut market is clearing. The seller of haircuts might want to sell more than he actually sells. If so, quantity supplied exceeds quantity sold, even if quantity sold equals quantity demanded. Especially when talking about money, we need to be doubly careful here. Our words confuse us.

2. More substantive. I think Kaldor and Trevithick are totally wrong. If people use extra money to pay down overdrafts, not just temporarily but permanently, this means that those same people must originally have had overdrafts that were higher than desired. They previously (before getting the extra money) had an excess demand for money, which has now been satisfied. The question is not "what would they do with extra money?" but "what would they have done otherwise, if they had not got that extra money?". It is the effect of the extra money *relative to what they would have done otherwise* that matters. If they would otherwise have paid off that overdraft by buying fewer goods than they sold, and the extra money means they continue to spend as much as they earn, this means the extra money has had a hot potato effect.

Update: put it another way: that extra money has eliminated what would have been a negative hot potato effect from those people paying down overdrafts.

Yep. A lot in common. Those guys understand the question. They understand why it is a very important question (even before "QE"). They are asking the right question. (Their answers might be wrong, but that's less important!)

Perry Mehrling is one of those people I should read more. If you have any particular suggestion, I would be grateful.

Take a very simple and really extreme real-world example: Zimbabwe. The growth in the stock of money was determined by the (Robert Mugabe's) demand for loans from the central bank, at the (very low) real rate of interest the bank set. But RM didn't want to *hold* that money he borrowed - he wanted to *spend* it. And the people who got that extra money didn't want to hold it either; they (eventually) spent it, and it never got returned to the central bank (who would be stupid enough to do that with it, at very low real interest rates?). And that hot potato process bid up prices, which increased expected inflation, which made the hot potatoes even hotter and increased their velocity of circulation.

What's more, RM could probably repay all of his borrowing from the central bank just by giving it an apple or two.

The beauty of extreme examples is you can ignore all the ceteris paribus stuff. It's too small to matter.

Canada is exactly like Zimbabwe, except the central bank adjusts the interest rate in response to changes in the demand for loans and changes in the excess demand for money. The Bank of Canada's interest rate is not exogenous with respect to the stock of money created.

"The beauty of extreme examples is you can ignore all the ceteris paribus stuff. It's too small to matter."

I like that quote. Some people just don't get the beauty of extreme examples. :D

Nick, what would it take for a cell of fanatical, yet highly disciplined and secretive NGDPLT terrorists, working secretly within key positions (but not leadership positions) in the government and central bank to carry out their twisted vision and successfully do 5% NGDPLT? Say one of them had the keys to the printing press room, and could sneak in after hours to print up some unauthorized reserve notes, and another one's day job was to be responsible for "accurately" recording when reserve note liabilities were returned to the central bank. What's the minimum sized cell of zealots required, and what would they ideally need to have as official day jobs? (BTW, these sickos are so brain washed they don't care at all about the consequences of their actions, other than that NGDPLT is maintained at 5%, ... while the official leadership of the CB is blindly doing whatever it is they are doing (say like whatever the Fed is doing now) totally unaware of this cell of traitors in their mix! What would be those unintended consequences? What damage would these terrorists do?

The comments to this entry are closed.

• WWW