The orthodox New Keynesian position on liquidity preference and loanable funds

I am not an orthodox New Keynesian macroeconomist (ONKM), but I can pretend to be one.

Q: What determines the rate of interest?

ONKM: "The central bank sets the rate of interest."

Discussion: the above answer is a pure liquidity preference theory of the rate of interest. By having a perfectly elastic money supply curve, at some rate of interest chosen by the central bank, the stock of money adjusts to equal whatever quantity  of money is demanded at that rate of interest. Like in all liquidity preference theories, the rate of interest is determined by the demand for money and the supply of money. The only difference here is that the money supply curve is perfectly interest-elastic.

Q: But what determines where the central bank chooses to set the rate of interest?

ONKM: "Loanable funds."

Discussion: this is the bit that needs some explanation. Because the ONKM won't actually say the words "loanable funds". What the ONKM will actually say is something like:

ONKM: "The central bank chooses to set the rate of interest that it believes is compatible with keeping output at potential and inflation on target."

So we need to translate that answer into loanable funds language:

Let output demanded (call it Yd) be a negative function of the rate of interest r, a positive function of actual income Y, and a function of other stuff X.

Yd = D(r,Y,X)

And the ONKM central bank wants to set r such that output demanded equals potential output Y*, so that:

D(r,Y*,X) = Y*

Assume a closed economy for simplicity, subtract Cd (consumption demand) plus Gd (government demand) from both sides, remember the accounting identities C+I+G=Y and S=Y-C-G, where I is investment and S is national saving, and we get:

Id(r,Y*,X) = Sd(r,Y*,X)

The central bank sets a rate of interest such that desired investment at potential output equals desired national saving at potential output. Which is precisely the loanable funds theory of the rate of interest.

An inflation-targeting central bank will set a rate of interest equal to the rate of interest predicted by the loanable funds theory.

One problem with the above is that the central bank does not observe potential output Y*, nor does it observe all the elements in the set X, and it does not know the function D(.) either. It has to make estimates/guesses of all these things. So it would be more correct to say that the central bank sets the rate of interest where it thinks the loanable funds theory says it will be. So the loanable funds theory will only be true up to a random(?) error caused by the central bank's mistakes in hitting its inflation target.

Any ONKM will immediately acknowledge that problem.

A second problem is that there is not just one rate of interest but a whole slew of interest rates. Differing by indexed or not, by term, by risk, by liquidity, etc.

Any ONKM will immediately acknowledge that problem.

Setting those problems aside, I have my own disagreements with the ONKM perspective on this question. But this post is not about my own views.

This post is about Lars Syll's views. I would like to ask Lars if he agrees or disagrees with the ONKM view, as I have presented it/translated it above. Because a lot of stuff really does get lost in translation sometimes.

[Update: Here is Lars' response, and my comments on his response.]

[My guess is that both Paul Krugman and Greg Mankiw would roughly agree with the above, but I could be wrong.]

P.S. Astute readers may have noticed that I have said nothing about either "the natural rate of interest" or "the money multiplier" in this post. That's because they are irrelevant to this question. Whether or not there exists an inflation-invariant natural real interest rate depends on the functional form of D(r,Y,X). If we interpret r as the nominal rate of interest, then an inflation-invariant natural real rate of interest requires the derivative of D wrt r always equal minus the derivative of D wrt expected inflation. (Expected inflation will be one element in X.) Whether or not there exists a stable "money multiplier" depends on whether or not commercial banks expand and contract their balance sheets in exact proportion to expansions and contractions in the central bank's balance sheet. Neither affects the loanable funds vs liquidity preference question.

Comments

[Gaffers]

I definitely can't speak for Lars, but I think most post-Keynesians would disagree with the ONKM view that you've presented (which I really like, by the way, because it's so simple and clear). I think PKs strongly endorse the answer to the first question (endogenous money), i.e. that Central Banks set the benchmark interest rate. Where they disagree is with respect to the way that the investment and saving functions are specified.

They would disagree with the investment function because either (i) investment is not a negative function of the interest rate, due to the results of the Cambridge Capital Controversies (there are no well-behaved demand curves for capital); or (ii) the X component of the investment function is extremely volatile (reflecting animal spirits), such that the central bank will not be able to accurately target the natural rate of interest r* to offset changes in the X component. This is especially true if the economy is in a liquidity trap or is stuck at the zero-lower bound (which are different situations, by the way), such that r is sticky. So the PK investment function might look like this:

I = Id(Y,X)

According to PKs, fundamental uncertainty and environment-consistent rationality (as opposed to model-consistent rationality, i.e. rational expectations) mean that people do not optimise intertemporally with respect to their saving and consumption decisions, and instead make decisions based on rules of thumb. A basic old Keynesian savings function based on current income kind of captures this idea (although maybe Duesenberry's relative income hypothesis is better, empirically):

S = s(Y)

In other words, the rate of interest cannot equalise investment and saving. There is no market for loanable funds, so to speak. Investment and saving can only be equalised in a closed economy with a neutral fiscal stance by changes in output. Investment is considered the exogenous variable because of the answer to the first question, i.e. firms can finance their investments with money created by banks; they aren't limited by the stock of money in existence.

That's not to say that aggregate demand isn't a function of the rate of interest. Consumption demand (Cd) may be a function of the interest rate due to the income redistribution effects of interest rate changes and different propensities to consume out of different factor incomes. So in ISLM terms, the IS curve would still be downward sloping, but rather because consumption is a negative function of the interest rate. The LM curve would be horizontal, but you could still model a central bank reaction function, provided that there is some kind of Phillips Curve positive relationship between output and inflation.

[biagio bossone]

To Gaffers:
This is one of the most concise and effective representations of the PK position. Loanable funds theory has nothing to do with Keynes' theory of liquidity preference and income formation. Excellent comment. Many thanks!

[Nick Rowe]

Gaffers: and thanks for that clear comment. Yes, if Id and Sd were both independent of r (if the IS curve were vertical), then the loanable funds theory of r would not work.

But there is one thing in your comment that doesn't add up for me: if Cd is a negative function of r, for given Y, then Sd must be a positive function of r, for given Y.

[Luis Enrique]

Gaffers,

I am more convinced by empirical evidence than I am by Cambridge theorising that investment demand isn't very sensitive to interest rates (although there is some evidence that some firms react sometimes) however investment is sensitive to demand, and demand responds to interest rates via other channels, as you suggest, how important is the difference between PK and ONKM?

also I don't see the relevance of the 'money created by banks' bit - why do firms care how the money they borrow is created, they still have to pay an interest rate for borrowing it, it's not as if ONKM theory thinks the stock of money is the constraint on investment is it? afaik ONKM doesn't go there.

but these comments are not nearly as interesting as Nick's!

[Nick Rowe]

Luis Enrique: "also I don't see the relevance of the 'money created by banks' bit"

It is possible to make theoretical sense of an equation that goes something like:

Id = Sd + Msdot - Mddot

I had a go at it in a couple of my monetary disequilibrium posts. And maybe, just maybe, something like that is what the PKers too have at the back of their minds. (But now I'm wandering off onto my own view, rather than my interpretation of the ONKM view.)

[Ritwik]

A lot of the debate typically called liquidity preference vs. loanable funds is not so much about 'what determines the interest rate', as it is usually parsed, but about 'is the act of financing (typically, investment) contingent upon (presumably prior)saving'.

It's possible, though disappointingly uncommon, to have a 'loanable funds' theory of interest rates, while having a 'liquidity preference' model of financing.

[Gaffers]

Nick: thanks for your reply. Yes you’re correct, sorry I didn’t make that clearer when I considered the case of income redistribution. If Cd is a negative function of r, for given Y, then yes, Sd must be a positive function of r.

I suppose the key element to this PK model is that the investment function is independent of the rate of interest.

Luis: Thanks for pointing out the similarities between the PK and ONKM short-period models. Firstly, yes, I agree with you that investment seems to respond to demand. I think the investment function is best described with some sort of accelerator mechanism. There's been a fair bit of empirical stuff on that I think, and also some on the interest-inelasticity of investment demand. I can scrape around for links if you like?

Secondly, you’re quite right, the PK and ONKM models are ostensibly very similar; PKs and NKs would agree, for example, that the policy-mix for our current economic woes should involve low interest rates. Their disagreements arise partially because PKs don’t accept that microfoundations are necessary. Thus, they don’t consider Ricardian Equivalence to be relevant. The income distribution effects on Cd noted above also mean that incomes policies and other redistributive policies are on the table when it comes to managing demand.

I think the difference between NKs and PKs becomes more apparent in the long-run. PKs think growth is demand-led and endogenous, see e.g. the neo-Kaleckian growth models. They also think that capitalist economies generate endogenous business cycles, see e.g. Minsky’s financial instability hypothesis.

Thirdly, thanks for your question about the endogenous money comment. I phrased that very poorly. What I meant to say is that higher Id leads to higher Y which may lead to increased Md. If Ms is given, then higher Md will lead to a higher r, which will cause ‘crowding out’ if Yd is a negative function of r. If Ms is endogenous and adjusts to Md, there will be no crowding out. So yes, I think you may be onto something Nick!

Fourthly, I agree with you wholeheartedly that Nick’s comments are much more interesting (and thoughtful) than mine! I also really admire the clarity of his writing!

[Luis Enrique]

Thanks for that gracious and thorough reply Gaffer.

no need to dig around on my behalf!

I confess I don't really know what demand-led long-run growth means, I don't suppose we could just 'demand' our way to perpetual high growth without facing supply constraints ... is the mainstream story of long run growth driven by (endogenous) technological (broadly interpreted) progress rejected by PKers?

I hesitate to go near this debate again, but I thought the ONKM position was that both broad money via private credit creation and base money supply is endogenous under an interest rate targeting CB, although afaik mainstream macro doesn't model the details of that, perhaps to its loss I don't know. Plus I don't think many mainstream macro types think of Ricardian Equivalence as a real world phenomenon, more a theoretical result in a very stripped down model.

[Gaffers]

Luis: No worries! When people are curious about PK models, I try to explain with what limited knowledge I have.

I don’t want to derail the comments on this blog post, so I’ll just recommend the works of Marc Lavoie and Amitava Dutt on neo-Kaleckian growth (there are other PK growth theories, but this is their workhorse model). But basically, yes, they do say we can effectively ‘demand’ our way to higher growth under the assumption that there is always spare capacity in the economy (an odd assumption, I know, but it’s based on empirical evidence that most large firms operate at 80-90% of capacity at most times), and as long as investment (in growth terms) is a positive function of the rate of capacity utilisation. I’m not too sure about how they deal with technical progress in PK growth models, but I think that most subscribe to the ‘stylised fact’ of the Kaldor-Verdoorn growth law: that labour productivity depends positively on the growth rate. Also, due to ‘efficiency wage’ effects or the 'Webb effect', labour productivity also might depend positively on the level of real wages. PKs hesitate to discuss changes in TFP because of perceived theoretical problems with neoclassical production functions.

I’m not sure how ONKMs model money either, but on the face of it, it seems to be very similar to the PK theory of endogenous money. And yeah, if Ricardian Equivalence is dropped, then NKs get even closer to PKs in terms of short-run policy analysis.

[W. Peden]

Ritwik,

"It's possible, though disappointingly uncommon, to have a 'loanable funds' theory of interest rates, while having a 'liquidity preference' model of financing."

Very interesting. Would you like to say more about such a position? What broader consequences would it have?

[Nick Rowe]

Ritwik: "It's possible, though disappointingly uncommon, to have a 'loanable funds' theory of interest rates, while having a 'liquidity preference' model of financing."

I'm guessing: so, to take an extreme version: the central bank sets r to ensure that Saving = investment at potential output, but all saving takes the form of holding money rather than bonds or real assets? Makes sense. That's what's happening in my previous post on the optimum quantity of money.

Gaffers: "I suppose the key element to this PK model is that the investment function is independent of the rate of interest."

But some NK models (the very simplest ones) ignore investment altogether. But as long as C depends on r, they behave much the same way (except there's one less source of shocks, of course).

The way you get "crowding out" in NK models is a bit different. If G increases (temporarily), that would cause Y to increase above Y* if the central bank kept r constant. But the central bank wants to keep Y at Y*, so the central bank will increase r to prevent Y rising above Y*. This means that the central bank acts to make M actually *fall* if G increases. The central bank tries to make the LM curve vertical at Y*.

You write good comments. Stick around.

[Majromax]

As a bit of data, the relationship between investment share of GDP (and government+investment share) versus supposed "real" interest rates (nominal rates minus inflation expectation):

Two things are striking. The first is that while I is variable, G+I is much, much less so. The second is that while weak, I and r seem to have a positive correlation; increases in the believed real interest rate are associated with increases in private investment. The correlation appears to be stronger for the one-year government debt versus the corporate debt, but that may be a difference of durations (the inflation expectations are one-year).

[Dan Kervick]

Nick, I'm wondering about this passage:

Id(r,Y*,X) = Sd(r,Y*,X)

The central bank sets a rate of interest such that desired investment at potential output equals desired national saving at potential output. Which is precisely the loanable funds theory of the rate of interest.

I understand that Id = Sd is the equilibrium condition that determines the rate of interest in the loanable funds theory. But that doesn't mean that every theory that includes the same equilibrium condition is therefore a loanable funds theory, or that every possible account of the functional determinants of Id and Sd is consistent with loanable funds theory. To get a loanable funds account, don't more have to be added about constraints on the funds market? Can't you have a theory that sat entials there is an equilibrium state in which Id = Sd, but which doesn't entail that funds saved are the source of funds invested?

[Dan. I just rescued this comment from the spam filter. You may be getting caught in spam on other typepad blogs. NR]

[Nick Rowe]

Majromax: trouble is, we don't know whether shifts in the S curve are causing movements along the I curve, or whether shifts in the I curve are causing movements along the S curve. Or both.

[David Beckworth]

Can we also frame this issue from a time-horizon perspective: in the short-run liquidity preference, in the long-run loanable funds?

[Nick Rowe]

David: we could maybe do that. And that is closer to the traditional ISLM perspective (short run with P fixed is a mixture of LP and LF, while long run with P flexible is pure LF).

But in Canada, the short run is only 6.5 weeks (the BoC re-sets the overnight rate 8 times per year, if it thinks it needs to) and nothing much happens macroeconomically in 6.5 weeks. Perhaps the more relevant short run/long run distinction is how long it takes the Bank of Canada to learn about shocks to Y* and X.

[Nick Rowe]

Dan Kervick: really sorry, but something went wrong with your comment. I saw it in the spam filter, tried to fish it out, but now it is nowhere. Maybe I screwed up. Sorry again.

[Nick Rowe]

Dan: Aha! Your comment has reappeared!

Most (all?) macro theories will say that desired investment = desired saving. But it is the loanable funds theory that says it is the rate of interest that adjusts to equilibrate the two. If we said that it is real income Y that adjusts to equilibrate Id and Sd, that is not loanable funds. For example, if I wrote down:

Id = Sd(Y) then it is Y that must adjust, and that is not loanable funds.

The key for loanable funds is that Y stays at Y*, and it is r that adjusts to keep Id and Sd equal.

[Nick Rowe]

The simplest ISLM model, for example, is a mix of both liquidity preference and loanable funds, because *both* Y and r adjust to equilibrate *both* Ms=Md and Id=Sd.

The equations are: Ms=Md(r,Y) and Id(r)=Sd(Y)

[anon]

David Beckworth, Nick: keep in mind that we're always in the "long run" wrt. choices that occurred some time ago. Thus, it could be argued that LF explains current interest rates just as much as LP does, just in a different perspective. Also it's not clear why we should identify the "P flexible" time horizon as the one where "pure LF" starts to apply; indeed, there are things that suggest this is not quite the case.

[Dan Kervick]

Thanks for finding my comment Nick.

But it is the loanable funds theory that says it is the rate of interest that adjusts to equilibrate the two.

Doesn't it say more than that? If desired savings and desired investment at time t1 are each determined separately by r (along with some additional stuff X), and are equilibrated under the right conditions, but if the actual quantity of investment at t1 is not constrained by the actual quantity of savings at some earlier time t0, then it doesn't seem to me that the r-determined equilibrium has anything to do with funds that are "loanable", and that the intutive idea inherent to the loanable funds theory that the funds for investment are provided by the savings of economic units has been lost.

I guess this goes back to my old complaint that economic theories - in their present form at least - are not just mathematical formalism, but consist in formalism supplemented by causal narratives, usually expressed in ordinary language, that entail causal claims and information that is not captured by the formalism alone.

It might be possible to fully formalize these theories, but the formalism has to be extended to include symbols explicitly representing causal relationships. These relationships cannot be read off from the acausal langauge of simultaneous equations in equilibrium theories. This seems unbelievably important when considering policy interventions.

[Also retrieved from spam. Maybe if you post a couple more comments (anything is OK) and I retrieve them, typepad will learn? NR]

[Gaffers]

Nick: You make an interesting point about “crowding out” in NK models. It seems like a very strange way to use the expression “crowding out” though, since it requires policy intervention that, as you noted above, involves a fairly inexact, though educated, guess on the part of central banks as to the value of r* that corresponds with Y*. This seems different to the “crowding out” in classical and old Keynesian/old monetarist models, where any change in autonomous expenditures has an automatic effect on r (unless of course we are in the horizontal part of the LM curve of the OK model, i.e. a liquidity trap).

The PK model would look much the same as the NK model if you have the central bank effectively setting r to target Y*. But PKs emphasise that central banks may not be able to do this for a number of reasons, including the volatility of the X component of Id. Therefore, the ‘somewhat comprehensive socialisation of investment’ that Keynes talked about becomes useful for reducing that volatility and enabling more exact fiscal and monetary policy interventions to achieve Y*.

What interests me, though, is what happens in the NK model when there is no central bank reaction function so that the LM curve is horizontal. I think you posted on this before when you said that central banks in the NK model try to make the economy look classical, when in reality it is anything but. But on top of this, what would happen if Y has a unit root, so that Y* is endogenous? Then we lose long-run neutrality of money and the gap between NKs and PKs basically collapses. Of course, there is much more to the PK research programme than long-run non-neutrality of money, but I think that ONKM + endogenous growth is much closer to some PK models than most people realise.

[Nick Edmonds]

Nick,

This is great post, precisely because (as you say) a lot of this gets lost in translation.

For example the Lars Syll post you link to has a long quote explaining why a decision by individuals to save more does not automatically induce more investment by somehow freeing up available funds. But that really isn't a feature of NK models.

[Nick Rowe]

anon: "..keep in mind that we're always in the "long run" wrt. choices that occurred some time ago. Thus, it could be argued that LF explains current interest rates just as much as LP does, just in a different perspective."

OK. Sort of.

"Also it's not clear why we should identify the "P flexible" time horizon as the one where "pure LF" starts to apply; indeed, there are things that suggest this is not quite the case."

I'm not following you here. Do you mean if labour supply is a positive function of r?

Dan: I'm not sure I'm following you. So I'm going to say 3 semi-random things in the hope one of them is useful:

1. Actual saving is always equal to actual investment, by definition. Like apples actually bought = apples actually sold, by definition. That will be true in all theories.

2. Yes, it's not enough to talk about the equilibrium condition. We also need some sort of story of how we get to that equilibrium, if it shifts.

3. To my mind (I'm wandering off the NK reservation here) the key distinction is between saving in the form of extra money, and all other forms of saving. Take a very simple model of a monetary economy where money is currency printed by the central bank. I get my income in the form of money. There are 4 things I can do with my money income: buy consumption goods; buy investment goods; lend it to someone else; simply add it to the money already in my pocket, and leave it there. Those last 3 things are all "saving". But only the 4th thing is dangerous (unless the person I lend it to in the 3rd thing does the 4th thing and simply leaves it in his pocket). If I start to do the 4th thing (unless someone else does the negative version of the 4th thing), investment will not rise, and the economy will go into recession. So the central bank must print enough new money to stop me trying to do that 4th thing.

[Nick Rowe]

Gaffers: "It seems like a very strange way to use the expression “crowding out” though, since it requires policy intervention that, as you noted above, involves a fairly inexact, though educated, guess on the part of central banks as to the value of r* that corresponds with Y*. This seems different to the “crowding out” in classical and old Keynesian/old monetarist models, where any change in autonomous expenditures has an automatic effect on r (unless of course we are in the horizontal part of the LM curve of the OK model, i.e. a liquidity trap)."

Correct. It is very different. But I would quarrel with your use of the words "automatic" and "policy intervention". This comes down to my old question: what do we mean by the central bank "doing nothing"? Does that mean holding r constant? Does it mean holding M constant? The price of gold constant? The exchange rate constant? Inflation constant? NGDP constant? There are 1,001 different ways of "doing nothing". And all it really means is doing what we think of as "normal".

"What interests me, though, is what happens in the NK model when there is no central bank reaction function so that the LM curve is horizontal."

In an NK model, if the central bank holds r fixed forever, never responding to what is happening in the economy, the economy will either explode into hyperinflation, or implode into hyperdeflation. Either way the monetary system and central bank will be destroyed, and someone somewhere will create a new money. It's not a feasible policy.

What happens when we take an NK model, and add an AK growth model to it? Yep, we get a unit root in Y, because Y* will be permanently affected by past central bank mistakes. (There is a stock/flow problem here, because you need a higher *level* of K to give a higher level of Y*, but central bank mistakes only affect the *growth* of K. So Y must inevitably outrun Y* initially. But let's set that aside.) The problem is that this unit root exists, but cannot be exploited by the central bank. To exploit it, the central bank would need to make "mistakes" with an upside bias, so that Y is bigger than Y* on average. But given the NK Phillips Curve, that would require actual inflation to be above expected inflation on average. And given even the weakest and most reasonable version of rational expectations, that can't happen.

[Nick Rowe]

Nick E: thanks! I'm really hoping more PKs will read this, so that we at least get better communication on this stuff.

One of the interesting things, to me, is that the PKs always talk about the "funding" question when they talk about LF. And NK's talk about I and S. If I go off the NK reservation again, I think that PK's are trying to say something important here, but they aren't saying it at all clearly. Just like Steve Keen. And I think it all has to do with disequilibrium money. And it's similar to what I've been trying to say, in some old posts. Neither LF nor LP are true, except under very implausible assumptions about knowledge. We are usually "off" both the IS and LM curves. There's an Old Monetarist hot potato where Ms > Md, and an Old Keynesian slow multiplier where Id > Sd, and these are the same thing.

[Nick Rowe]

I said "And NK's talk about I and S."

Which isn't strictly true. NK's talk about Y, C, and I. You have to infer what they are saying about S.

[Jussi]

Great discussion! I hope Lars will follow-up.

[elatedcitizen]

thank you N.Rowe - to me, all this is really educational

(thumbs up for Gaffers)

[Steve Roth]

@Nick: "3. lend it to someone else; 4. simply add it to the money already in my pocket, and leave it there"

This is where things get dicey for me.

If I put the money in the bank, I'm lending it to someone else. (3)

So it is only in the case of people literally hoarding more physical cash in mattresses (4) that "investment will not rise, and the economy will go into recession"?

Given the on-demand interchangeability of cash and reserves, does this make sense?

This relates to what seems a central misconception to me: that *not* transferring $10K from my bank account to your bank account ("saving" in both the vernacular and NIPA "personal saving" senses) will result in there being more aggregate "savings" (i.e. loanable funds) than if I did transfer $10K from my bank account to your bank account ("spend" -- say, to pay you as my personal masseuse).

There's the same amount of money in household-sector bank accounts either way. "Personal saving" (i.e. the vernacular sense, "money saving") doesn't increase the aggregate quantity of "loanable funds". It just reduces the extant savings' velocity WRT purchases of newly-produced goods and services.

[Jacques René Giguère]

Steve: if you put money in a easy to withdraw form, the bank is less likely to lend it. No different than a mattress.

[Frank Restly]

Steve,

"If I put the money in the bank, I'm lending it to someone else."

But the bank must find a borrower for those funds otherwise the bank is serving as your mattress.

Couple other things I can do with money:

(4) I can pay wages to workers who use tools to change the look / functionality of an existing good (iron ore to iron to steel)
(5) I can burn / destroy it
(6) I can give it away (charity)
(7) I can wear it around my neck (Veblen good) - not exactly the same as stuffing it in the mattress because of the underlying motivations - I wear it around my neck to show off my "wealth" and encourage envy, I stuff it in my mattress to hide my "wealth" to discourage envy. I would imagine that different "levels of envy" in an economy have different time paths.

[Nick Rowe]

Steve: "If I put the money in the bank, I'm lending it to someone else. (3)"

Nope. That depends on what the bank does next, which depends on what the central bank does next. Plus, there are now two different types of money: central bank money and commercial bank money, and if you put it in your chequing account it's still in your (metaphorical) pocket.

That way lies a total ball of confusion. Simplify, simplify, simplify. Start out be assuming One Big Bank, that is both a central bank and a commercial bank. That issues only one type of money. And it does not matter if that money is paper or electrons. Now make an assumption about what the Bank holds constant: is it r, M, or NGDP, or what? Then ask your question.

The rest of your comment now makes sense. If the One big Bank holds M constant, then if we all try to increase our stock of M by spending less, we obviously must fail in aggregate. The only thing that happens when we try to save more in the form of money is that V falls, so PY must fall.

[Nick Rowe]

Jussi and elated: thanks!

[Oliver]

Hi Nick, I'll join in with the praise for your post.

And I'll allow myself to join in the comments although I fear I'm far below par for this course.

You write:

There are 4 things I can do with my money income: buy consumption goods; buy investment goods; lend it to someone else; simply add it to the money already in my pocket, and leave it there. Those last 3 things are all "saving". But only the 4th thing is dangerous (unless the person I lend it to in the 3rd thing does the 4th thing and simply leaves it in his pocket). If I start to do the 4th thing (unless someone else does the negative version of the 4th thing), investment will not rise, and the economy will go into recession.

First, why do you assume investment will automatically rise if any of 1-3 occurs? In case 1, all that has happened is that there is now one consumption good less on earth, but still the same amount of money around. How does that induce investment? In 2, real and financial saving have just exchanged hands. Absolutely nothing has happened. In 3, either 1, 2 or 4 can happen in future. But as of now, nothing has happened, except that spread sheets have become longer by 4 entries in all.

As for

Steve: "If I put the money in the bank, I'm lending it to someone else. (3)"

Nope. That depends on what the bank does next, which depends on what the central bank does next. Plus, there are now two different types of money: central bank money and commercial bank money, and if you put it in your chequing account it's still in your (metaphorical) pocket.

For the bank's books to balance, something must happen immediately. One of two things can happen immediately. a) either the bank increases its asset side, say by extending a loan (new investment), or b) it decreases some other liability by the same amount, say by writing down its equity or paying down debt (no overall change in investment). What happens next, is only of interest insofar as it is causally linked to what happens now. So, in what way do a) or b) depend on what the central bank does next seeing as firstly they happen before the central bank does anything and secondly are opposite wrt to their effect on investment? And quite frankly, banks can only ever be interested in a) in the medium term, no matter what the CB does. b) = certain death.

[Nick Rowe]

Oliver: "First, why do you assume investment will automatically rise if any of 1-3 occurs?"

I don't. Only if 2 occurs does my investment automatically rise with my saving.

"In 2, real and financial saving have just exchanged hands."

Don't even think of making a distinction between "real and financial saving". Because next you will end up in some sort of Marxian BS like "real vs financial capital".

"Saving" is "income minus taxes minus consumption", and I have ignored taxes here.

And don't even think about spreadsheets. Once we get accounting involved, everything gets even more confused!

[Steve Roth]

@Jacques: "if you put money in a easy to withdraw form, the bank is less likely to lend it."

My impression is that consumer deposits are considered -- by both banks and regulators -- to be among the most reliable and stable funding sources. I think you're confusing the individual and the aggregate here.

[Oliver]

@ Nick

First I'll have to correct my golf analogy. Below par is a good thing in golf. But anyway...

I don't. Only if 2 occurs does my investment automatically rise with my saving.

OK. In that case I don't know what investment will not rise, and the economy will go into recession. means. Do you mean cannot rise because the money that would be necessary for funding further investment cannot be used because it's in my pocket? In that case, I posit that at least PKers - not sure about NKers (maybe implicitly?) - would deny that constraint exists.

And as for Only if 2 occurs does my investment automatically rise with my saving, do you mean overall investment? Because I'd say that it is indeterminate whether overall investment rises when I buy an investment good. Just as it's indeterminate when I buy a consumption good or do nothing. Of course, if everybody wants to do the same thing, something has to give. And the difference in all these models seems to be what that something is or ought to be.

Re Marx: no class warfare intended. But I'll admit sloppy thinking on my behalf. I'll rephrase that to a title to an existing investment good changed hands.

Not sure what you mean by your spreadsheet comment. I'd say accounting clarifies many things because balance sheets have implications. Two people with the same net worth but otherwise completely different balance sheet compositions will face very different consequences when something on the outside changes. Same goes for the economy as a whole, I'd say. There's much more meat there than in an interest rate set by the CB, for example.

[Dan Kervick]

1. Actual saving is always equal to actual investment, by definition. Like apples actually bought = apples actually sold, by definition. That will be true in all theories.

Only in the long term, at best, no? What if a given society, in a given year t and starting with a stock of wealth W, receives an income Y from work down in year t-1. Suppose they consume Y/2 and they also save Y/2 by laying it up in a storehouse. Suppose none of their existing wealth is lost due to waste or depreciation. But suppose they also engage in no productive activity whatsover aimed at yielding income either in year t or any year later than t. Then their actual saving in year t will be Y/2, their new level of wealth will be W + Y/2 and their investment will be zero.

[Nick Rowe]

Dan: "Only in the long term, at best, no?"

No. Every single minute. "To the penny!" (as MMTers are fond of saying.)

In your example, the wheat they put in the storehouse is included in "investment". It's "inventory investment". Yes, I know....It's a tautology.

[Dan Kervick]

That's a really bad definition of investment, then. The wheat in the storehouse isn't a business inventory of inputs intended for use in future production. It's just an inventory of consumables meant for future consumption.

In this case, we don't even have an informative macroeconomic identity - just a redefinition of "investment" to mean something like "net addition to wealth" or in other words a pure synonym of "saving" rather than anything having to do with the employment of resources in production.

[Jerry Brown]

Dan Kervick, in your scenario of a society, how do they have any income at all, let alone savings from that, if they engage in no productive activity at all? From some other society? And are you talking about saving in financial assets or in real physical goods?

[Nick Rowe]

Dan: "...just a redefinition of "investment" to mean something like "net addition to wealth" or in other words a pure synonym of "saving" rather than anything having to do with the employment of resources in production."

Yep! "Investment" is a synonym for "saving", just like "apples bought" is a synonym for "apples sold". I tried to explain this to the MMTers, who think that I=S (and its open economy variant) reveals the deepest truth about the universe, and who wander around chanting it like some hare Krishna mantra.

But are any identities informative, except about how we use words?

[Oliver]

MTers, who think that I=S (and its open economy variant) reveals the deepest truth about the universe, and who wander around chanting it like some hare Krishna mantra.

From my understanding, I think it is actually what they are accusing others of doing. E.g. when theses others say things such as:

The central bank sets a rate of interest such that desired investment at potential output equals desired national saving at potential output.

followed by:

"Investment" is a synonym for "saving"

Put both together and you get The central bank sets a rate of interest such that desired national saving at potential output equals desired national saving at potential output.. Which isn't surprising because they are identities, after all.

So some then venture out to look for ways to separate saving / investment with a real world counterpart from that without - only to be derided as communist pigs.

I'm happy to be convinced that that isn't the way to go. But I'm not particularly convinced by what I feel is an attempt to circumvent tautologies by adding mystic qualifiers such as 'desired' in the hope that one can thereby decouple one side of the equation from the other.

And just to be fair, I think MMTers and PKers make the same mistake when they talk about 'effective' demand. Since supply = demand just as S = I, it makes no sense to me to say that one follows the other in any meaningful sense. One cannot construct equilibrium conditions out of either because they are always in equilibrium, by definition.

[Nick Rowe]

Oliver: apples bought is the same as apples sold. But apples (some) people **want** to buy (quantity demanded) is not the same as apples (other) people **want** to sell (quantity supplied). I want to sell you 10 apples but you want me to sell you only 8 apples. The price of apples adjusts until the desires of different people are mutually consistent.

Same with *desired* saving and *desired* investment. The rate of interest adjusts (or output adjusts) until the desires of different people are mutually consistent.

[Nick Rowe]

And the big difference comes with saving in the form of money -- the medium of exchange.

If I want to increase my saving in the form of land, I need someone else's consent -- because someone has to sell me land.

If I want to increase my saving in the form of bonds, I need someone else's consent -- because someone has to sell me bonds.

But if I want to increase my saving in the form of money, I don't need anyone else's consent -- I just spend less money.

[Lars Syll]

Nick, sorry for not having had time to respond earlier, but I just put up a post on my blog in response to your loanable funds musings: http://larspsyll.wordpress.com/2014/09/21/the-loanable-funds-fallacy/

[Nick Rowe]

Thanks Lars! And I have left 2 comments there in response.

[Oliver]

I see his satanic majesty has spoken ;-), so I guess you'll have more interesting things to do than engage with a troll like me. But I'm still struggling to understand, so i'll keep posting.

apples bought is the same as apples sold. But apples (some) people **want** to buy (quantity demanded) is not the same as apples (other) people **want** to sell (quantity supplied). I want to sell you 10 apples but you want me to sell you only 8 apples. The price of apples adjusts until the desires of different people are mutually consistent.

At the price at which apples were sold, quantity demanded by consumer / investor was identical to quantity willingly supplied by producer.

One can watch prices over time and see whether they're moving. But to deduce from any such movement that demand isn't keeping up with supply or vice versa, or that person x wanted to sell 10 apples but couldn't, putting the tautology question aside for a minute, seems a bit wild.

The fact that the value of my Rothko has been rising cannot be due to the fact that the market wants two of it. Maybe the market just believes art prices will go up long enough for participants to make a buck? Nifty investors could just be borrowing new money and buying art. Works 'til it doesn't.

Nor must prices on all other goods fall for the price of my Rothko to rise, which is what I think loanable funds is trying to make us believe.

I'm completely lost with your second comment.

Once on has land, bonds or money, it's easy to save in either of the three, right? So you're saying it's easier to get ones hands on money than on land or bonds?

Why? I could inherit land. And even if so, why is it important?

[Oliver]

I see you provided your answer to my last question in your second response to Lars Syll. I'll think it through.

[Oliver]

Before I attempt to do so: am I correct in assuming that you were speaking as Nick Rowe the Monetarist, as opposed to Nick Rowe the ONKM, in your second response (to me, above)?

[Nick Rowe]

Oliver: Yes. When I start rabbiting on about how money is really different, that's the real me speaking.

[Too Much Fed]

Can there be 5 quantities?

quantity demanded

quantity bought

quantity sold

quantity supplied

quantity potential

[Nick Rowe]

TMF: No. That's only 4. Quantity bought and quantity sold are the same thing.

[JKH]

Nick,

totally OT, but reminded me in a weird way of things you've written about the reserve system:

http://monetaryrealism.com/the-full-monty-on-naked-short-selling/

[J.V. Dubois]

Nick @Sep 19, 04:49PM: Only if 2 occurs does my investment automatically rise with my saving.

This is technically correct since you talk about "my" but I think that it may hide a little bit of truth hre. If saving is income minus consumption AND Saving = Investment (which it does) then investment "automatically" occurs whenever you have income and do not spend money on consumption. Even if you hold that income in your purse. Andy Harless explained very succinctly in his post here: http://blog.andyharless.com/2009/11/investment-makes-saving-possible.html

So what happens when you receive a cash income and you put it into the box? The thing is that you received the income in exchange for something. The person that paid you either bought investment or consumption goods from you. If it is consumption that was bought from you, then the buyer dis-saved money (he spend money he had saved in his purse for consumption goods) but since he gave money to you, and you by definition "saved" that exact amount the second after you recieved the payment - the net saving is unchanged. If buyer bought investment goods from you then he "invested" (as opposed to consumed) and you saved in equal equal amount. So both: saving and investment increased.

I understand that this is nothing new and it tells nothing about overall income over some time periods or anything else except that S=I by definition. But I think it is a useful exercise for everybody who things investment is something else - people who talk about financial capital, wealth and whatnot inevitably confusing discussions. I know that Andy's article cleared that for me quite well.

[Oliver]

There is nothing to prevent an individual manicurist from saving more, in the form of money. He just buys fewer haircuts. So there will inevitably be a recession (unless the central bank supplies more money).

It is saving *in the form of money* — the medium of exchange — that causes the problem. The medium of exchange is very very different from other assets like land.

What you are describing is the paradox of thrift, which goes back to Keynes (possibly further, I don't know). But in any case, I doubt any Keynesian, New, Post or otherwise would deny the existence of the phenomenon.

Tying in with my point above, though: there is also nothing, save taxes (which applies to the manicurist, too) or expropriation, to prevent a house owner from continuing to own his house (save). So there must be a similar phenomenon to the paradox of thrift wrt to land or bonds or whatever. An asset market freeze-up? A liquidity trap?

But I agree with you that there must also be a significant difference. Here's my take on what that is:
The manicurist, through not spending his income, is preventing something new, output (in this case a haircut) from being produced. Income = output. A recession is defined as a drop in output. It is an interruption of the flow of newly produced goods for immediate or gradual consumption (investment).

In an aset price recession, it is only already existing things that are not changing hands. There is no income and no output involved.

That is also why I would say that the term 'printing money' is too much of a simplification to properly address this issue. And it has nothing to do with the fact that money comes in bits and bytes nowadays. My beef is that it only captures one half of the transaction. The other half - what is it printed for? - is lost.

Helicopters do not drop money from the sky. New money, and I'm referring to bank money here, not CB money, is 'printed' explicitly for a purpose. It can either be directed at providing income by financing output, thus tackling the problem of a recession directly or at proppoing up prices of existing assets, addressing a recession indirectly, at best (via the wealth effect).

That is also why I would say that accounting is a helpful tool, because it reminds us to always think of both sides of a transaction. It is a mental crutch for more consistency in our fariy tales.

And maybe the term *medium of exchange* is confusing in this context, too. Asset swaps also involve the medium of exchange, but they do not produce recessions.

[Oliver]

re: quantity demanded

quantity bought / quantity sold

quantity supplied

quantity potential

Matias Vernengo seems to think it's all woo-woo (in the comments):

http://nakedkeynesianism.blogspot.ch/2011/11/neo-wicksellian-macroeconomics.html

Actually no. No need for any discussion of desired and actual I and S.

[Oliver]

A loving couple have a history of exchanging favours. Husband H likes giving his wife manicures, whereas Wife W likes giving his husband haircuts.

To keep track of each others favours over time, they decide to introduce a simple monetary economy.

They agree initially that 1 haircut = 1 manicure.

In the first preiod, they each perform one haircut / manicure, respectively, issuing an IOU to the other each time.

So, we can say: output = 2; consumption = 2; money extant = 2

Now, in period 2, assuming they always take turns, we have two choices:

economy 1) W & H can redeem their respective IOUs.

economy 2) W & H can do the same as in period 1 and issue new IOUs for their respective services.

So in period 2, we can say: output = 2, consumption = 2, money extant is either 0 or 2.

Which economy is more prone to recessions? Does the amount of money matter? And in what way? Does adding an interest rate change things? Does it matter whether the IOUs are evenly distributed?

We can imagine further that one day W finds out H has been having an affair, so she withholds her haircuts in response to which H withholds his manicures. We have a recession.

Enter marriage counsellor CB. She diagnoses that it is an acute lack of money that is causing the recession. So, while nobody is watching, she flies her toy helicopter into H & W's house and drops a bunch of IOUs into each of their wallets. Will that cure the recession?

If I were the counsellor, I'd personally convince H (lend him money) to perform a bunch of manicures on W as an apology and hope that that restores confidence. It ain't the money, it's the manicures.

[Nick Rowe]

JV and Oliver:

If I want to buy a newly produced machine, but can't find a seller, then my Id > I.

If I want to sell a newly-produced machine, but can't find a buyer, so have to hold it myself, then my Id < I.

If I want to buy land, but can't find a seller, then my Sd < S.

If I want to sell land, but can't find a buyer, then my Sd > S.

But money is different, because it is flowing both into and out of our pockets. I can always reduce my inventory of money, by buying less money or selling more money. Nobody can stop me buying less money. I can always increase my inventory of money, by buying more money or selling less money. Nobody can stop me selling less money. But in aggregate, we can't all do that.

Andy talks about money a bit. Matias just says "Actually no. No need for any discussion of desired and actual I and S." without explaining why he thinks that. We do need a discussion, and we need to distinguish money from all other goods.

[Steve Roth]

@Nick: "is it r, M, or NGDP, or what? Then ask your question."

Been pondering, not sure if you're still following here...

I get this. It's taken me years...

But:

I think that thinking only makes sense given the assumption I question in the second part of my comment: that monetary saving by real-sector entities increases the *stock* of monetary savings. (Back to my -- and your -- old bugaboo word...) No matter how many times I go around with IS-LM, that assumption seems to be at its core.

And I don't think that's true, don't think it can be true.

I'm talking about "personal saving" here as defined in the NIPAs -- disposable income less expenditures. (DI being income +/- taxes/transfers to/from guv sector).

As for "savings," the only measure I can think of for that is household net worth (with firms' value imputed to households as shareholders).

[J.V. Dubois]

Nick: I get what you are saying and I kind of agree. I say "kind of" because technically you are incorrect as it is impossible to increase your money inventory without selling something else for money first. You may stop purchasing anything and live only from home production. But to actually get some money you just have to sell something for it. It will not magically appear in your wallet. And at the moment you do it you have an income that you saved automatically the nanosecond after you earned it.

- If you sold consumption goods for money then net saving (and investment) did not change. Money were dis-saved from wallet of your buyer but they are now in saved in your wallet in
equal amount.

- If you sold investment goods for money then your buyer increased part of his income that was not spent on consumption (increased his saving). He also dis-saved equal ammount of money from his wallet but similar to previous example that money was just transfered and saved by you. This increased aggregate saving is equal to amount of investment goods that were purchased. S=I.

[Nick Rowe]

JV: you are right, of course. What I said only works for someone who already has a flow of money coming into his pocket, as well as a flow out. If everyone cuts their flow out in half, it is possible that some people will see their flow in be cut to zero. They are unable to buy money any money at all. I was considering a small change, starting from an initial equilibrium where everyone has a flow in and a flow out.

When the representative agent reduces his flow out, he is surprised to discover that his flow in falls by exactly the same amount, so he has not increased his stock of money at all. I say "surprised", because the representative agent does not know he is the representative agent. Plus, even if he did know, it's a Nash equilibrium, where he takes others' actions as given.

[Ritwik]

W.Peden

Thanks, not quite sure what all the implications of holding the position I described are - best guess is you end up somewhere between Leijonhufvud amd Mehrling. more Leijonhufvud than Mehrling.

[Jussi]

@Rowe: 'Matias just says "Actually no. No need for any discussion of desired and actual I and S." without explaining why he thinks that. We do need a discussion, and we need to distinguish money from all other goods.'

Matias followed-up and explained: "--The actual I and S are not equal". Not sure what he means by 'actual' but he might be mistaken here.

[Nick Rowe]

Jussi: I just read Matias' comment. I too think he is mistaken. Read literally, his comment makes no sense. But I think his mistake can be fixed. This is exactly the same as the problem Steve Keen ran into, which can also be fixed.

[Oliver]

Matias has responded:

http://nakedkeynesianism.blogspot.ch/2011/11/neo-wicksellian-macroeconomics.html

Not sure whether this reflects Matias' position, but I think I would agree with this:

http://www.csbancari.ch/pubblicazioni/RMElab/gnos.pdf

[Nick Rowe]

Thanks Oliver: I find Matias' answer unsatisfactory. If each agent's choice depends on the choices of other agents, and all agents move simultaneously, and makes his choice without knowing what the others are choosing, what ensures we get to the Nash equilibrium? (Unless we have common knowledge and all agents can solve the model.) If something changes, and agents make different choices, they will all be surprised by others' choices, and will regret their own choices (unless they all have perfect information and can solve the model).

I found the paper by Claude Gnos unhelpful.

[Oliver]

I don't want to beat a dead thread, nor plow your front lawn, but I'm still a bit confused. So if you'll excuse me, I'll try and dumb this down to my own level:

At all measurable points in time, I=S.

From one point in time to the next, I=S may change, so I1=S1 =/= I2=S2.

Payments are points in time.

Changes in I=S over time are due, on the one hand, to to the fact that people don't always do the the same thing did before, both individually and in aggregate. They change their minds for various reasons that are subject to studies by psychologists, historians, economists, theologists, biologists etc... Individuals may see an overall trend and conclude that although their own actions change, overall outcomes won't. They're often wrong. Furthermore, the ideas people have now may not turn out to be as good as they want them to in future.

Thus, neither individual nor aggregate expected outomes must equal actual outcomes. Actual outcomes can exceeded or fall behind expected outcomes. Nevertheless, for each actual outcome, S=I. There is never a measurable ex ante point in time where S=/=I.

Payments can be explaind in exact accounting terms, all else can't. We need a story. Nash equilibrium may be one of those stories of how we get from a to b, although I very much doubt it's sufficient to account for the complexity of humanity.

I understand Matias as saying that the main determinants of the discrepancy between aggregate expected and aggregate actual outcomes is when saving/investment decisions do not translate 1:1 into the productivity development of capital. This can be plotted as trend curves with an intersection portraying actual outcomes.

Personally, for someone who likes to cry 'radical uncertainty', such curves send the wrong message. But anyway...

The balancing item between nominal investment/saving and actual productivity of capital is the market value of assets invested in. So, if I borrow new bank money to invest in a new piece of machinery but nobody buys its produce, the value of overall investment = aggregate saving remains the same, but the value of my asset (the machine) will decline. Ultimately to 0, leaving the economy with a debt that it cannot repay.

This can happen if the products of the machine are effectively bad. But also if, for whatever reason, people decide not to spend out of their income, i.e. they don't consume. Collectively, such action will render the initial investments unprofitable.

This is the point that Gnos is stressing (I think). He seems to be saying that distribution of income must be such that the spending power of labour keeps up with productivity of capital. Effective demand is a function of the propensity to consume out of income. That propensity insures that investment remains profitable.

Which part(s) do you not agree with?

[Nick Rowe]

Oliver: apples bought and apples sold are just two different ways of describing exactly the same number. We cannot say that apples bought determines apples sold, or apples sold determines apples bought, because they are the same thing.

Same with actual saving and actual investment.

You can't even draw two different curves.

None of this helps us explain what determines apples-bought-and-sold, or saving-and-investment.

"This can be plotted as trend curves with an intersection portraying actual outcomes."

What are the two curves? What is on the axes?

[Oliver]

Hey, that is what I've been trying to say :-)! (Matias is the one arguing with curves, not me...).

I thought you were saying that there was a state of the world in which S =/= I.

Can we agree that nobody is saying that but that we've falsely been accusing each other of doing so?

[Nick Rowe]

Oliver: "Hey, that is what I've been trying to say :-)!"

Yep. And I was repeating it, in my own words, to ensure we are on the same page!

There is a state of the world in which desired/planned/expected S =/= desired/planned/expected I.

And what happens when that happens tells us what determines I (or S, same thing).

[Oliver]

OK, great! We can team up on Matias :-).

Re your second sentence:

Are you looking for a tool by which to equilibrate expectations about S and I?

If so, I think that's an oxymoron. We cannot have differing expectations about one and the same thing. Hopes, maybe, but that's moving into religion...

If the economy plans to invest 1 billion next year, it cannot expect to save more or less (money) than that amount.

That doesn't change if you break it down.

If investors decide to invest more than consumers want to save, consumers will consume more, leaving investors to either save their own investment or disinvest (pay down debt). In the latter case, there is no saving / investment over the whole period.

And in either case, I don't think you can say that there is a disequilibrium of aggregate saving and investment desires at any point.

What I think one can say though is, that there are more or less desirable absolute levels of investment & consumption for an economy. But I would say that is important to look at them in terms of flows, not in terms of money stocks accumulated. Flows determine activity levels, which determine employment, whereas the stocks are just residuals of saving desires. Hence my insistence on the importance of income as opposed to the money stock. Of course, a money stock can be considered potential for future activity, but then so can an overdraft facility.