Keynes, New Keynesians, and the Keynesian Cross

Simplify massively. Ignore investment, government spending and taxes, and exports and imports. All output is consumed.

Assume a Keynesian consumption function: Cd = a + bY. Draw Samuelson's "Keynesian Cross" diagram. Real consumption demand Cd on the vertical axis, and real output (real income) Y on the horizontal axis. Demand for consumption this current period depends on realised sales of output this current period. "Equilibrium" is where the consumption function crosses the 45 degree line. Because that tells us the level of output at which demand for output equals that same level of output. But that "equilibrium" is not necessarily a full demand = supply equilibrium. Output may be less than "full employment" equilibrium. Firms and households may want to produce and sell more output and labour than they are able to sell, if demand is too low. The multiplier (the change in "equilibrium" output divided by the upward shift in the consumption function that caused it) is 1/(1-b). There is positive feedback in this model, because demand depends on itself.

What does that Keynesian Cross diagram look like in a New Keynesian model, with infinitely-lived consumption-smoothing agents who never face binding borrowing constraints? That depends on how long the "current period" is, and on whether income in the current period affects expected future income.

Assume that the representative agent is perfectly confident that his income will be equal to "full employment" income in all future periods, regardless of what happens to current income. (Yes, this assumption is very questionable.)

If the current period is very short, the consumption function will be almost horizontal (the marginal propensity to consume "b" will be almost zero). Because changes in current income have a negligible effect on permanent income. And if the central bank reduces the current period real interest rate, the consumption function will shift up, causing an increase in "equilibrium" output. If the central bank sets the right interest rate, so the real interest rate equals the "natural rate", the intercept of the consumption function (autonomous consumption "a") will be almost equal to full employment income.

Now let's make the current period longer and longer. So that current income has a bigger and bigger effect on permanent income. Autonomous consumption (the intercept) gets smaller and smaller, and the marginal propensity to consume (the slope) gets bigger and bigger. Less obviously, a given-sized cut in the real interest rate will cause a smaller and smaller upward shift in the consumption function.

Now let's take this lengthening of the current period to the limit. Suppose the current period lasts forever. So my bolded questionable assumption above becomes redundant, because there is no future period. Let us put permanent income on the horizontal axis, and permanent consumption on the vertical axis. Autonomous consumption drops to zero, the marginal propensity to consume rises to one, and the effect of a cut in the real interest rate drops to zero. The budget constraint (plus non-satiation) ensures that permanent consumption must equal permanent income, for any level of permanent income, and for any real interest rate. This means that the consumption function necessarily coincides with the 45 degree line.

And this means that any level of permanent income, between zero and full employment income, is an "equilibrium". Animal spirits are the only thing that determine which of those equilibria will be chosen. Since b=1, the multiplier would be infinite, if it were possible to shift the consumption function up. But the budget constraint says it is not possible to shift the consumption function up, so that infinite multiplier is purely hypothetical.

Now let us introduce a stock of money into the model. The representative agent holds a stock of money. He has a desired stock of money, and an actual stock of money, and the two stocks may not be the same. Introducing money into the model breaks the identity between permanent consumption and permanent income. If an agent holds more money than he wishes to hold, he will plan to run down his stock of money over time, by spending more than his income, so that his permanent consumption will be greater than his permanent income. And if an agent holds less money than he wishes to hold, he will plan to increase his stock of money over time, by spending less than his income, so that his permanent consumption will be less than his permanent income. But even though each individual agent can add to or subtract from his stock of money, by spending less or more than his income, this is not possible in aggregate.

If the desired stock of money is independent of permanent income, an increase in the actual stock of money, so that it is greater than the desired stock, will cause an upward shift in the consumption function. It does not matter if that upward shift is very small, because b=1 and so the multiplier is infinitely big. An extra $1 hot potato in circulation will be enough to expand the economy from 50% unemployment to hit the full employment constraint. And withdrawing a single $1 from circulation would cause the economy to collapse to 100% unemployment.

But a more plausible assumption would be that the desired stock of money is an increasing function of permanent income. This ensures that the consumption function has a slope of slightly less than one, so we get a determinate "equilibrium" level of permanent income where the desired stock of money equals the actual stock.

The model would be more plausible still if we assumed that the desired stock of money depends on permanent nominal income. There is a real stock of money at which equilibrium permanent income is equal to full employment permanent income. Maybe (or maybe not, if it causes destabilising expectations) a slow adjustment of prices, in response to output being greater or less than full employment, would eventually bring the economy to full employment equilibrium.

Keynes understood that demand depended on itself. General Theory chapter 3. Keynes understood that the only thing that might (or might not) bring the economy to full employment income was some sort of real balance effect on aggregate demand (and that the central bank could do the same thing quicker and better than waiting for prices to adjust to get the right level of real money balances). General Theory chapter 17. New Keynesians have forgotten Keynes' two insights about aggregate demand. They have forgotten that demand depends on itself. They have forgotten that some sort of real balance effect is the only thing that might (or might not) bring the economy to full employment automatically. New Keynesians are guilty of an intertemporal fallacy of composition: even if the right rate of interest will ensure full employment in any single period, it does not follow that the right rate of interest will ensure full employment in all periods. New Keynesians just assume, without any explanation, without anything in their models that would make this the only rational expectation, that agents expect that future income will equal full employment income. New Keynesians are the "classical" economists who just assume (the agents in their models expect) an automatic tendency towards full employment. New Keynesians have models of monetary exchange without money. New Keynesians are not Keynesians.

Just continuing my lovely argument with Roger Farmer about Keynes and modern macro.

Comments

[JKH]

"autonomous consumption (the intercept) gets smaller and smaller"

What's making that happen?

[Nick Rowe]

JKH: Suppose your income was 0 for t years, and 100 per year thereafter forever. "Permanent income" is defined as the annuitised value of your income from today till the end of time. The longer is t, the less is your permanent income, so the less you consume, because consumption depends on permanent income, and on the rate of interest, and your "permanent consumption" (the annuitised value of consumption from today till the end of time) equals your permanent income.

[JKH]

I understand that.

But not in context. Why does increasing current income as you describe mean autonomous consumption goes to zero?

It seems like an important point to be very clear on, given that it's essential (to the geometry) for the rest of your thesis. Maybe it's obvious to everybody else, but I'm not seeing it.

[Nick Rowe]

JKH: You misread me. Autonomous consumption means "how much the representative agent would want to consume if his current income were zero". We are not increasing current income. We are increasing the length of the "current period". If the current period of zero income lasts for 2 years, rather than one year, before we get back to full employment, that will cause current consumption to drop. And increasing the "current period" to 3, 4, 5, etc. years, will cause it to drop more and more. And if the "current period" lasts forever, autonomous consumption would drop to 0.

[JW Mason]

JKH- Autonomous consumption here simply means the consumption that is based on future periods' income, as opposed to the current period's income. So the longer the current period is, the smaller the share of income coming from future periods, and the lower is autonomous spending.

[JW Mason]

In this model there is no autonomous consumption in the Old Keynesian sense, consumption unrelated to income.

[Nick Rowe]

An increase in current income will always increase current consumption, by an amount "b". If the "current period" is very short, b will be very small, and near zero. If the "current period" is very long, b will be large, and near one.

[Nick Edmonds]

Agree with all this, except I don't think it needs specifically to be money. The same applies in relation to any financial asset, provided an increased desire to hold that asset doesn't imply an increased desire to provide funding to someone who is credit constrained. So you could replace the term "money" with "safe assets" say and your analysis should apply. (And telling the story using "safe assets" rather than "money" is probably closer to the situation of the last few years).

[Nick Rowe]

JW: "In this model there is no autonomous consumption in the Old Keynesian sense, consumption unrelated to income."

I would say that in the NK model (PIH), and with expected future income exogenous, if the current period is very short, (almost) all current consumption will be autonomous in the sense of being (almost) unrelated to current income.

[JKH]

"Autonomous consumption here simply means the consumption that is based on future periods' income, as opposed to the current period's income. So the longer the current period is, the smaller the share of income coming from future periods, and the lower is autonomous spending."

Now that would explain it clearly for me if that is generally understood.

This is the 'traditional' KC right, where a = autonomous consumption?

Maybe its basic stuff, but I hadn't realized that this definition was specified as time period dependent in precisely that way.

[JW Mason]

Nick-

As usual, I see no problems with the logic of your argument. I agree with it as a critique of New Keynesian models. And I agree, you can find this argument clearly stated in the GT.

But, this is not the only way to resolve the problem, and you can find other resolutions suggested in the GT too.

At the point where you introduce the money stock, there's a different option. You can consider everything up to that point as a proof by contradiction of the permanent income hypothesis. Rational expectations means, strictly speaking, that a model should assume that agents' belief about the future values of a variable are the same as the values the model itself predicts. In this case, the model makes no predictions about the future level of income. So we can't describe people as acting on the basis of its predictions.

At this point, we have to fall back on the kind of behavioral story that Keynes used. It is for whatever reason a "psychological law" that expenditures move with income, but less than one for one. (We could come up with reasons.) It will still always turn out to be the case ex post that that aggregate income equals aggregate expenditure. But it will not be the case for any individual unit. There is no such thing as an intertemporal budget constraint.

In this story, there is no automatic mechanism bringing the economy to full employment. (Altho there is a tendency for the economy to return to whatever level of output is deemed normal, provided the disturbance was not large and/or long enough to change agents' views of what is normal.) Changes in interest rates or in the money supply are just one instrument among others for changing the flow of current expenditure.

(I think this is basically Roger Farmer's position. But he feels professionally obligated to describe this kind of indeterminacy as being consistent with rational expectations.)

[JW Mason]

DISproof by contradiction

[JW Mason]

Nick Edmonds-

People often say this but I don't think it's right. Excess demand or supply for safe assets can be eliminated by a change in their price. There is no reason to think it requires income adjustment. Price adjustment doesn't work for money because of its special role as the unit of account.

[JW Mason]

Incidentally, Keynes himself does make a version of the safe asset argument made by Nick Edmonds above, suggesting that historically land -- as the safe asset -- might have depressed demand for currently produced goods and services in the same way that demand for liquidity in the form of money has in more recent times. But as Samuelson pointed out (in "Land and the Rate of Interest"), this was a mistake on Keynes' part: "Keynes erred in likening land to money: he can be charged with forgetting that the price of land is a variable that can be bid up to whatever level is needed to bring the interest rate down indefinitely close to zero." Samuelson goes on to say that Keynes was right to think that a preference for land as a store elf wealth could indeed depress growth, because investment in land competes with investment in fixed capital. But that effect doesn't show up in the simple model we are discussing here.

[JW Mason]

JKH-

It isn't. In the normal presentation, we have expenditure as a function of (a) current income and (b) everything else. Autonomous expenditure is the everything else part. Nick is saying that IF you apply this in a New Keynesian framework where agents are exactly satisfying an intertemporal budget constraint, then "everything else" reduces to future income.

[Nick Edmonds]

JW

I think that comes down to elasticities. A change in the price (yield) on bonds might have very little impact on the overall demand for safe assets, but may have a big impact on what proportion of those safe assets people want to hold as bonds. In that case a change in the price of bonds may eliminate the excess demand or supply of money.

[Nick Rowe]

Keynes was wrong on land. Only money works. Back later.

[Ralph Musgrave]

“An extra $1 hot potato in circulation will be enough to expand the economy……” More or less what MMTers have been saying for some time. To use MMT parlance, a deficit increases “private sector net financial assets”, which tends to increase demand.

[JW Mason]

Nick-

I don't think so. If there is excess demand for bonds, the price of bonds will rise. The elasticity of demand for bonds tells you how big the price rise will be, but it doesn't provide any reason for the rise to stop while there is still excess demand. Since we observe safe assets being held at a positive yield, we can be sure that excess demand for safe assets is not holding back expenditure.

[JW Mason]

Keynes was wrong on land. Only money works.

Right, in this model. If we introduce investment in fixed capital, then the existence of land can depress output.

More or less what MMTers have been saying for some time.

Yes, MMT (at least the older versions) and monetarism are equivalent. The only difference is whether you call the the money-emitting entity a central bank or a budget authority.

[JW Mason]

To be clear -- I agree with Nick R. and Samuelson that Keynes was wrong on this point. But -- as Samuelson also says -- there is a related point where he was right.

[Nick Edmonds]

Look at it another way. Say there are only two safe assets - money and bonds. And we have a household that has $200 of safe assets - $100 in money and $100 in bonds. But they would prefer to have $300 in safe assets, with the extra $100 in bonds, because they don't need more liquidity.

So we might say there's an excess demand for bonds, but it's not going to be one that moves the bond price, because they can't as yet bid for the extra bonds. We could call it an excess demand for money, but that's a bit odd because they don't actually want any more money, except in that they want more money so they can buy bonds. And if we tried to put more money into the system by lowering bond yields, so that they're happy to switch to $50 in bonds and $150 in money, say, it doesn't necessarily mean they don't still want an extra $100 in safe assets.

[Steve Roth]

@Nick Edmonds and @JW: Banging my spoon on the high chair again: I think you're talking past each other cause you're doing the old thing of confuting money with currency and close equivalents, conceptually.

If money is, rather, the exchange value embodied in financial assets (including currency and everything else), then Nick Edmond's statement:

"I don't think it needs specifically to be money. The same applies in relation to any financial asset."

Is true, prima facie.

And if we consider deeds to be financial assets (claims on real assets, though without an offsetting balance sheet liability on anothers' books) -- embodiments of money -- I think that also defuses the Samuelson objection to the Keynes argument. Or at least partially unties the gordian knot they're twisted into.

[JW Mason]

it's not going to be one that moves the bond price, because they can't as yet bid for the extra bonds.

I don't understand this. Why not? Everyone offers unsafe assets for sale in exchange for the safe ones, until relative prices change sufficiently that they are satisfied holding the current stock.

[Andy Harless]

I don't think New Keynesians "just assume (the agents in their models expect) an automatic tendency towards full employment." What they assume is that the agents in their models expect reasonable price level behavior in the long run, where "reasonable" means anything except hyperinflation or hyperdeflation. If you drop this assumption, there are an infinity of other equilibria for output but, if I'm not mistaken, only two other equilibria for the long run price level: it either approaches zero or it approaches infinity. Perhaps New Keynesians should regard these equilibria as theoretically possible, but that doesn't mean they must regard them as interesting. Hyperdeflation is something we have never seen in history, and it just seems insane that people might expect money to become infinitely valuable in the long run. (OK, I've snuck money into this argument, and you can say this is "something like a real balance effect," but it's surely not very much like a real balance effect. I'd think of it more as a temporary concession to the real world. IRL we know that money exists, and that fact is important for ruling out hyperdeflation, but otherwise, from a NK view, not very important and can be ignored in the model.) As for hyperinflation, I believe it can be ruled out by allowing the central bank to specify a reaction function in which it can, in principle, deviate from the "right" interest rate. Basically, "If y'all start to expect hyperinflation, we're going to do something crazy to make sure it doesn't happen." But Chuck Norris doesn't actually have to beat anyone up, and the actual path of the interest rate doesn't have to deviate from the "right" path.

[JW Mason]

Steve, I don't think that works. Again, financial assets' prices can change. And they do change, very easily. I don't think the idea of excess demand for financial assets in general makes sense.

To make this kind of argument work, you need to consider the liability side of balance sheets as well as the asset side. What units want is not liquid or safe assets, but a safer or more liquid balance sheet position. This includes demand for less debt. But the price of outstanding debt cannot change, it is fixed in nominal terms. And if units devote more of their expenditure to paying down debt and less to currently produced goods and services, the price adjustment runs in the wrong direction. Nor can units with excess liquidity transfer it to the units with excess demand for liquidity, except to the extent that the latter hold salable assets. But by definition, a unit with excess demand for liquidity must hold only relatively illiquid assets, so netting them out against debts will be costly and slow.

This is basically the balance-sheet recession idea (often, but incorrectly, attributed to Richard Koo -- it goes back at least 30 years, probably more). It is the logical extension of Keynes' liquidity preference story of aggregate demand as the inverse of money demand, for a modern financial economy.

This story works logically, unlike the safe asset shortage story. That doesn't mean it describes what's going on in the US or elsewhere right now. I don't think it does. I think what we have now is precisely the situation Nick describes before he brings money in. Spending is low, so incomes are low, so expected incomes are low, so desired spending is low. No excess demand for anything.

[Kevin Donoghue]

Nick: "New Keynesians have models of monetary exchange without money. New Keynesians are not Keynesians."

You'll never settle the question, who are the true heirs of Keynes? He left umpteen Last-Will-&-Testaments, so to speak, without saying which one was really his last word. In The Crisis in Keynesian Economics John Hicks describes a model which is every bit as objectionable (from your point of view) as the NK model. Hicks calls it Keynes's wage-theorem. He writes: "It is because of the theorem that investment, and income, and money supply are measured in wage-units; for it follows from the theorem that when so measured, they are independent of the level of money wages."

He sums up: "the view which emerges from the General Theory...is nothing less than the view that inflation does not matter." To me, that looks remarkably like the NK model (or Jordi Gali's version at least, since that's the one I happen to have wrestled with). Keynes really wasn't as bothered by nominal indeterminacy as you are; not always, anyhow.

(I think Hicks was wrong to call this endogenous-money wage-theorem notion the view which emerges from the GT; it's just one of the many possibilities Keynes envisaged.)

[Min]

"Demand for consumption this current period depends on realised sales of output this current period."

Ah! The time travel assumption. Demand for the next time period depends on current sales.

"'Equilibrium' is where the consumption function crosses the 45 degree line."

Equilibrium only makes sense if we are talking about a time lag, so that future demand equals current demand (and future sales equals current sales). Because:

"There is positive feedback in this model,"

Without a time difference there is no feedback.

"because demand depends on itself."

Because future demand depends on current demand.

[Nick Edmonds]

JW

"Everyone offers unsafe assets for sale in exchange for the safe ones, until relative prices change sufficiently that they are satisfied holding the current stock."

That assumes that they want to reduce holdings of unsafe assets / increase holdings of safe assets. What I was meant was when they want to reduce consumption / increase holdings of safe assets, which I think is the parallel to what Nick is saying. Otherwise if people wanted to reduce holdings of unsafe assets / increase holdings of money, presumably what you are saying would also apply, i.e all that would happen would that the price of unsafe assets would change.

That's probably my fault for talking about safe assets. I only did so, because I wanted to distinguish the case where increased appetite for assets led to some credit constrained person increasing consumption. But I can see it's confusing, because demand for safe assets is sometimes discussed in the context of a switch out of unsafe assets. I'd tend to agree that that argument is more problematic.

[JW Mason]

Nick, to say there is excess demand for X means that at current prices, you would like to exchange more not-X for X.

What I was meant was when they want to reduce consumption / increase holdings of safe assets

Right. You still have not explained why this does not silly lead to a rise in the price of safe assets.

That's probably my fault for talking about safe assets. I only did so, because I wanted to distinguish the case where increased appetite for assets led to some credit constrained person increasing consumption.

If you didn't mean safe assets, why did you say safe assets? And what did you really mean?

Anyway, we have not even gotten to the argument you make here -- that even if asset prices could not adjust, non-monetary assets (and, for that matter, money as well) are produced by the private sector. So why doesn't an excess demand for assets of whatever kind lead to increased issuance of such assets, i.e. increased borrowing? Your notion here is that there is only increased demand for safe assets, and that issuers of safe assets will not issue more, for some unspecified reason. So in this story, asset prices adjust, there is no excess demand for safe assets from asset owners, and income is not being held down by people's inability to realize their desired asset position. However, the shift in asset demand changes the interest rates faced by different borrowers. And for some reason, spending by safe borrowers is less responsive to interest costs than is spending by risky borrowers.

This story is logically coherent. But it's quite different from the model Nick is presenting here. You cannot just do a find-and-replace of "money" with "safe assets".

It's also not supported empirically, at least for the US. Risk spreads are have narrowed, not widened, and there is no evidence that credit constraints on consumption or investment are tighter than pre 2008.

[JW Mason]

(simply, not silly. I should turn off autocorrect.)

[JW Mason]

I should add: I think the flight from risky assets and resulting credit-rationing is a good story for the crisis period itself. But not for the slow recovery following it.

In general, I think it's more in the spirit of Keynes to say you need to talk about money/finance to explain transitions between high and low-level equilibria; but not to explain the existence of a low level equilibrium, once you are in one.

[Nick Rowe]

Introduce a fixed stock of land into the model. Suppose there is an excess demand for land. Would that cause desired permanent consumption to fall below permanent income?

Either the price of land is perfectly flexible or it is not perfectly flexible.

If the price of land is perfectly flexible, the price of land jumps up instantly, to eliminate the excess demand for land.

If the price of land is not perfectly flexible, there is an excess demand for land, but people will not be able to buy land, because there aren't enough willing sellers. Unable to buy land, people buy consumption instead.

Either way, an excess demand for land cannot cause desired permanent consumption to drop below permanent income.

For "land" read "bonds", or "whatever it is that MMT guys are talking about, except money".

Money is different. It is the medium of account, so its price is sticky. And it is the medium of exchange. There are two ways an individual can satisfy his excess demand for money: sell more goods; buy fewer goods. Nobody can stop him buying fewer goods. There is only one way to get more land: buy more land. If there are no willing sellers of land, you can't buy more land.

[Nick Rowe]

Kevin: "Keynes really wasn't as bothered by nominal indeterminacy as you are; not always, anyhow."

But what we have here is real indeterminacy.

Andy: we can't rule out nominal indeterminacy (hyperinflations or deflations) unless we can rule real indeterminacy. Given a Calvo Phillips Curve.

[Nick Rowe]

JW: Interesting what you say about Samuelson and land. Yep, if the price of land is flexible, keynes is wrong. But even if the price of land is stuck, it is still wrong that an excess demand for land can cause a recession. There is an excess demand for unobtainium too, but that doesn't cause a recession. The bit about land reducing investment and growth is the old "Junker Fallacy", except it isn't a fallacy in the OLG model.

[Kevin Donoghue]

Nick to Andy: "we can't rule out nominal indeterminacy (hyperinflations or deflations) unless we can rule real indeterminacy."

AFAICT the point Andy is making is that if we rule out hyperinflation & deflation, the real variables become determinate. I'm no expert but I'd guess he's right. Simon Wren-Lewis says much the same thing.

As for Keynes, he was fussy about some things but I don't think he'd have been too bothered about the assumption that Classical results emerge in the long run.

BTW, has anyone ever documented an instance of this alleged Junker Fallacy which crops up from time to time? Tyler Cowen gave some vague reference to Fritz Machlup years ago but I've never been able to locate the guilty party. Is it one of these "said nobody, ever" things?

[JW Mason]

Nick R.-

Agree on all counts. The only thing I would add is that there are also two ways to reduce your liabilities: sell more, and buy less. So in this sense borrowing less is just like increasing your money holdings. But with the difference, there is not a fixed stock of not-being-in-debt in the way that there can be a fixed stock of money.

[Steve Roth]

@JW:"I don't think the idea of excess demand for financial assets in general makes sense."

I couldn't agree more. But then I don't think the idea of excess demand for money in general makes sense. I think you may have asked in the past: if there's excess demand for money, why aren't people maxing out their credit cards?

You have the answer:

"What units want is not liquid or safe assets, but a safer or more liquid balance sheet position. This includes demand for less debt. "

Isn't that safer balance sheet comprised of both -- safer assets and less debt? Both balance sheet moves reduce the stock of exchange value as embodied in financial assets (the stock of "money"). The former through trading of price-volatile equities for less-volatile financial assets that have fixed payouts, the latter through debt payoff. That big balance-sheet shrink is far more rapid and far greater than is readily corrected via monetary or fiscal moves.

Those with limited assets can only shrink their balance sheets by spending less (on real goods). So while V might rise algebraically as M contracts, illiquidity and debt repayment is fighting against that.

It doesn't seem like demand-for-money or safe assets adds much understanding to all that.

[Tom Brown]

Steve, Cullen Roche has essentially asked this of Nick before:

"I couldn't agree more. But then I don't think the idea of excess demand for money in general makes sense. I think you may have asked in the past: if there's excess demand for money, why aren't people maxing out their credit cards?"

I think his response was that the credit card companies, because there's an excess demand for money, don't want to part with money (i.e. credit in this case) to purchase IOUs from people (i.e. there's a glut in the IOUs from people market): if you view getting the lending process as exchanging IOUs (with one of the IOUs being "money"). I probably butchered Nick's explanation, and perhaps it doesn't address the particulars of credit cards... but I thought I'd throw that out there.

[Tom Brown]

I'm following the above discussion with Nick Rowe, Nick Edmonds, Steve Roth and JW Mason with great interest, because I think it gets close to answering a question I have... but not quite. In a discussion with Nick Rowe elsewhere I've identified (I think!) a case where the quantity of MOA (M) can change, but this change has no long term effect on the average price level (P) because demand for MOA also changes. This happens because non-MOA (bank deposits) are exchanged for MOA (CB liabilities). The specific situation is a cashless society where a CB with $R of reserve-liabilities takes over the banking sector with $D deposit-liabilities, thus changing M by 100*(D-R)/R%, but not changing P at all. My question: are bank deposits the only non-MOA good this can happen with? What if the CB were to take over all money market funds and replace them with deposits at the CB? What about short term bond funds? The CB might pay a premium to make up for loss of interest in either case, or it could just pay interest. Can these also be examples of demand transference between non-MOA and MOA which eliminates any change in P?

[The Market Fiscalist]

"even if the right rate of interest will ensure full employment in any single period, it does not follow that the right rate of interest will ensure full employment in all periods"

I'm not sure I get the implications of this. If the single period in question is the current period, and in each period the interest rate is used to ensure full employment what different does it make if this doesn't ensure full employment in the future (you can set the appropriate rate for that period when you get there.)

Is the implications of this rather that if expectations aren't for full employment in the future that there may be no interest rate that will achieve full employment in the present either ?

[Nick Rowe]

TMF: that's the fallacy of composition. Each individual can see the stage better if he stands up (holding constant whether other people are sitting or standing), therefore (invalid inference) if every individual stands up they will all see the stage better.

Kevin: "AFAICT the point Andy is making is that if we rule out hyperinflation & deflation, the real variables become determinate. I'm no expert but I'd guess he's right. Simon Wren-Lewis says much the same thing."

He's right (for any reasonable Phillips Curve). But it's just a way to assume full employment by the back door.

[Roger Sparks]

Nick Rowe.

"Now let us introduce a stock of money into the model. ......"

At this point in the post, you introduce a second measure of trade into the Keynesian Cross. Now we have trade plus "stock of money". The volume of trade has increased because we have begun using "stock of money". It is easily seen that the exchange volume increase only occurred because not only goods were being exchanged, money was also exchanged.

As macroeconomist, we make a mistake when we only place money on the x and y axis of the Keynesian Cross. We should sum money and debt, clearly identify that we are doing that, and present our Keynesian Cross nomograph as a simple demonstration that income is always the same as expenditure WITH BOTH DEBT AND MONEY TRADED. The parties of the trade accepted both money and debt for labor and materials.

If we take this approach, we can see that the economy produced a result measured with last years data. The employment achieved was accomplished with a measured amount of DEBT AND MONEY. If even more employment is desired, we would need even more debt and money.

Of course we would argue about when debt BECOMES money. At least we would be recognizing that trade is not only for money; it also for debt.

[Nick Edmonds]

JW and Nick,

I feel that maybe we're making some different assumptions about what other things change respond when one thing changes.

I'm going to focus on bonds, specifically treasuries, rather than land. Let's say then that there is a money demand Md = m(Y,R), a demand for treasuries Td = t(Y,R) and a demand for other assets Ad = a(Y,R). R is the vector of yields.

So let's say that the function Td changes, so that Td increases, but the other functions stay the same. At current yields, people want to hold more treasuries, but no less money or other assets. So they try to do two things: a) they try to buy more bonds and b) they try to spend less on consumption.

I think this is where the difference is coming in. As far as I see it, they can't only try to do (a) because if they were successful that would result in a fall in their money holdings, and I have assumed that their money demand is unchanged. So they have to try and do (b) as well.

Then, of course, income falls so all the asset demands are potentially affected. What ultimately happens, depends on the elasticities in the functions but, in the extreme, the change in Y could entirely eliminate the excess demand for treasuries with no price changes at all. (It's worth noting that a fall in income may in practice increase the supply of treasuries as well, due to budget deficits, but that's not necessary for the argument.)

In your story about the land, Nick, people try to buy more land, whilst keeping consumption unchanged. They're only doing (a) not (b). This implies to me that they want to reduce their money balances. After all, if they were successful, that is what would happen - their money balances would decrease.

The corresponding position in your main post is where people want to hold more money, but they want to do so by reducing their treasury holdings. In that case, they would not cut their consumption. They would just try to sell treasuries and the yield on treasuries would rise. We might then find we end in a position where they decide that actually they're better off keeping their savings in treasuries after all, so they're happy sticking with the money holdings they've got.

I assumed when I read you post that this was not what you meant, but instead you meant that people wanted more money, whilst retaining the existing quantity of other assets. My point is that if people want to hold more treasuries, whilst retaining the same quantity of other assets (including money), then you get the same result as for money.

(The difference potentially arises when people wish to hold more of an asset that might fund someone who is credit-constrained, because that then potentially has impact on spending elsewhere. That was the reason for my careless use of the term "safe asset", but I think that was a bit of a red herring, so I don't think I should have used it in this context. Anyway, my main point, is that I don't think the effect is limited to money, but it really depends on what else you assume changes when the asset demand changes. If you assume that an increase in demand for money will be funded by reducing consumption, but that an increase in demand for other assets will be funded by running down money balances, then you will get different results.)

Does that sound right?

[Nick Rowe]

Nick E: start in equilibrium. Then all of a sudden each individual wants to consume less and hold more bonds/land/whatever.

This creates an excess demand for bonds/land/whatever. So the price of bonds/land/whatever rises (and the yield R on bonds/land/whatever falls) until people stop wanting to hold more bonds/land/whatever and stop wanting to consume less. We stay in equilibrium.

Unless the fall in yields on bond/land/whatever causes an excess demand for money.

JW: "Agree on all counts. The only thing I would add is that there are also two ways to reduce your liabilities: sell more, and buy less. So in this sense borrowing less is just like increasing your money holdings. But with the difference, there is not a fixed stock of not-being-in-debt in the way that there can be a fixed stock of money."

I'm still thinking about this. I'm thinking about red money.

BTW, OT: did you see the link I posted in the last comment on the 45 degree post. A real world supply-side multiplier in the Paul Krugman babysitter case. An increase in M causes Y to fall.

[Kevin Donoghue]

Nick,
Any model which tries to combine Keynesian and Classical features will have "a way to assume full employment by the back door", so I don't think that's a telling criticism. It all depends on how soundly the back door is constructed. In the case of the NK model (Gali version) it's done by assuming that (a) the interest-rate policy followed is sensible (it doesn't, for example, violate the Taylor Rule); and (b) agents draw comfort from this, in the sense that they don't foresee catastrophic inflation or deflation simply because these outcomes are mathematically possible.

If I'm right in saying that's the NK rationale, then I can't see much reason to object to it. If I'm wrong I'd appreciate you and/or Andy telling me.

[Nick Edmonds]

"Unless the fall in yields on bond/land/whatever causes an excess demand for money."

Aha. I think I see the source of disagreement. This statement I totally agree with. The effect on income of an increase demand for bonds depends, amongst other things, on the interest elasticity of demand for money. If demand for money is completely inelastic, then it will have no effect. (That's in the simply model. In the real world, and if we want to specify money as being the balance of transaction accounts, we also have to consider supply elasticities of money).

So our disagreement is not about whether a change in the demand can impact on consumption at all - it's a disagreement about likely parameter values.

[Nick Rowe]

Kevin: take my "model" above, once I introduce a stock of money, the demand for which depends on nominal permanent income. That model does not assume full employment via the back door. It goes in the front door. (It might miss the front door, if expectations are destabilising, but there is a front door.) Keynes talked about the front door to full employment in chapter 17, but said it would be easier for the economy to get in the front door if the central bank moved the door.

Nick E: I'm not sure we disagree on parameter values. I'm not sure we disagree at all. I want to force us to think in terms of X causing deficient AD only via the effect of X in causing an excess demand for money. And this forces us to think about whether a better monetary system/policy could prevent things like X causing deficient AD.

I insist on my perspective, because I think it is both theoretically correct, and policy-relevant. Monetary policies/systems are not written in stone. Things like X only cause deficient AD because bad monetary policy/systems let them.

[Nick Rowe]

Nick E: let me put it this way: when we talk about AD, we are really talking about the excess demand and excess supply of the medium of exchange. There is no such thing as "AD" in a non-monetary economy.

[Nick Rowe]

Plus, concentrating on the demand for newly-produced goods only is far too narrow. An excess demand for money causes a disruption in *all* monetary exchanges, not just exchanges of newly-produced goods. What happens in the markets for old houses, used cars, etc.? They get disrupted too, and they matter too. It's not just about AD=Cd+Id+Gd+NXd.

One old post.
A second old post.

[Nick Rowe]

A recession is *not* a fall in GDP. It is *not* a fall in economic activity. A recession is a fall in monetary exchange. Non-monetary activity rises in a recession. Home production, barter, etc, increases in a recession. A recession is a change in the *mix* of monetary vs non-monetary activity. We are just better at measuring the monetary activity. Plus, monetary exchange really is more efficient for a lot of stuff, so when we try to do it without money, we don't do it as well, and people are worse off. Unemployed growing their own veggies.

[Nick Edmonds]

"when we talk about AD, we are really talking about the excess demand and excess supply of the medium of exchange. There is no such thing as "AD" in a non-monetary economy."

I agree with the second statement, but I don't think it necessarily leads to the first. But I would agree that it may just come down to perspective.

I suppose the real test is not what you call it, but what might work. I can imagine a position where the private sector wants more B+M. You can call that an excess demand for M, but the question is can you fix it buying reducing B and increasing M. If you can't, I see that as an excess demand for (B+M), not an excess demand for M. I don't think it would ever be an excess demand for just B, even if it was caused by a change in the demand function for B.

[Nick Rowe]

Nick E: why stop at M+B? Why not add land in too? M+B+L. Why not add old houses, old cars, paintings, you name it? Where do you draw the line?

You (I'm guessing, like most "keynesian" economists) divide the world into two groups of things: newly-produced goods; everything else. (And what you have in mind as the archetype for "everything else" is "bonds".) Why divide the world up this way?

I divide the world into two groups of things: money; everything else. Money is different because everything else is bought and sold for money and priced in money.

This is about ontology. Borges celestial emporium of benevolent knowledge.

[JW Mason]

Nick-

Oh right, yes I saw the babysitting thing. That was GREAT. Perfect example. The only problem is the Mises guy didn't understand it, he said it was an example of how monetary stimulus leads to inflation. In fact, inflation is precisely what the babysitting economy needed and didn't get. Inflation would have brought home back to "full employment" in the same way that deflation is supposed to in the demand-constrained world, via the Keynes/Pigou/real balance effect.

[Kevin Donoghue]

Nick,
You're obviously entitled to assume an exogenous stock of money and a stable demand for real balances if you want to. That has its attractions. We end up with something much more like IS-LM. But I can understand why New Keynesians don't want to do that. For one thing, it's not a satisfactory way of modelling the actual behaviour of central banks. For another, demand-for-money functions don't fit very neatly into the Lucasian approach which NKs have opted for.

What I don't think you're entitled to say is that "New Keynesians are not Keynesians" when, in effect, all they are doing is making money endogenous. If the central bank is following a Taylor rule, M adjusts passively. Since it doesn't really add anything to the model we may as well leave it out. That's effectively what Keynes was doing in the passages which John Hicks found so unsettling.

[Steve Roth]

@Nick Rowe: "I'm still thinking about this. I'm thinking about red money."

Likewise. In particular I'm thinking about a question that's been vexing me for some time: is your untapped Visa credit limit "money"? Any help appreciated.

[JW Mason]

I don't have time to engage in the debate here properly right now. I might write a post on my own blog later. But two points I want to make now.

1. Nick R.'s argument is logically coherent and can be four in lots of postwar Keynesian as well as monetarist stuff. But it cannot be found in Keynes. The margin between current consumption and money holdings plays zero role in Keynes' account of how aggregate demand varies.

Keynes' story goes like this: current consumption, and therefore current saving, is strictly a function of income. (There is some discussion of changes in the propensity to consume in the GT, but in terms of secular trends, not business-cycle scale fluctuations.) With the level of saving fixed by income, savers (i.e. asset owners) have a choice between more and less liquid stores of wealth, conceived as money and bonds. Bond prices adjust to get equilibrium in this market, which determines the interest rate. The interest rate then determines the discount rate applied to future earnings from produced capital goods. This discount rate, **combined with variation in expectations about future income streams**, determines the demand price for capital goods; combined with the supply price, this determines the volume of new capital goods produced. Production of capital goods then determines income and closes the system.

In this story, nobody is reducing expenditure on non-money goods in order to acquire more money. All the action is in what Nick calls the "Junker fallacy."

2. Nick R. writes:

when we talk about AD, we are really talking about the excess demand and excess supply of the medium of exchange.

You say this a lot. So do other people. But I don't understand how you can be so insistent on this, when right here in this post you have described variation in AD that has nothing to do with excess supply or demand for the medium of exchange. Consider your NK model where the current period is very long -- that is, where people base their expectations about future income entirely on current income. In this economy, any level of activity is possible, depending on people's initial beliefs. If they believe there will be a high level of activity, they will be right, and if they believe there will be a low level of activity, they will also be right. In neither case need there be any excess supply or demand of the medium of exchange. This is also a form of AD, and as I emphasized above, it is the source of variation in AD that Keynes thought was most important in explaining fluctuations in real economies. Now, you are right that a fixed stock of money could help coordinate expectations on a particular level of income. But so could lots of other things.

[JW Mason]

On reflection, I'm wrong to say that Keynes' story is a form of the Junker fallacy. That would be true if we had wealth owners directly making the choice between capital goods and money, but that's not how Keynes tells the story. The distinction between rentiers and entrepreneurs is important to him -- nobody is making a choice to hold wealth in the form of fixed capital. Presumably the entrepreneurs operate with borrowed funds, and the banking system passively accommodates their desired borrowing. This is a bit of gap in Keynes story. But that doesn't change the fact that it's the story he tells.

[Odie]

"A recession is a fall in monetary exchange. Non-monetary activity rises in a recession. Home production, barter, etc, increases in a recession...Unemployed growing their own veggies."

I see where you are coming from but doubt it is really that clear cut. People will certainly try to substitute indispensable goods/services with their own labor but discretionary activity will also take a (big) hit. Not going on vacation, not attending college, not buying the new household appliance/furniture etc. Real losses of output that will not be substituted. Just look at the consumption of natural resources (oil, metals, lumber etc.); all indicators of real losses in economic activity not just substitution between monetary and non-monetary exchanges.

[Nick Edmonds]

"Where do you draw the line?"

I don't think there's an easy answer to that.

There is an important line between produced goods and the rest. But there's also an important line between fixed income monetary assets and the rest. And how important these lines are depend on elasticities, both for demand and supply. As I said, if the demand and supply of money are highly inelastic, then that might indeed be the only line that matters.

[Steve Roth]

@Nick Rowe: "I divide the world into two groups of things: money; everything else. Money is different because everything else is bought and sold for money and priced in money."

My highchair again: I divide the world into two groups of things: real assets (long-term "capital" and and short-term consumption goods), and financial assets. Financial assets are claims on real assets. They are not money (no, even currency isn't), they are embodiments of money -- exchange value. They cannot be consumed; they have no use value (but can be exchanged for things that do). They require no inputs to production (just lay down your credit card instead of your debit card and you've created money), and they are not inputs to production (they can only be exchanged for inputs to production). "Financial capital" is an oxymoron.

Money is different because it (and the financial assets that embody it) is produced for free, and can only be exchanged, not consumed.

What are the demand dynamics for a "commodity" that is produced with no inputs or cost, and that cannot be, is never, consumed? Are they identical to the dynamics for things that do require inputs, and are consumed? That's the (a) crux I've been wrestling with.

[JW Mason]

oh RIGHT, now I remember. This was another problem Keynes acknowledged after the publication of the GT. (I can't find the exact cite right now.) In the GT, he implicitly was considering a period of time long enough for capital projects to be carried to completion. Then the income resulting from investment would have generated savings equal to the value of the investment, presumably held in the form of equities, so that we don't need to consider any change in asset-market equilibrium. But in a later article, he acknowledged that the financing needs of investment while it was being carried out would create additional demand for liquidity, and that this factor probably explained a large part of interest rate movements over the business cycle.

[Kevin Donoghue]

If wanting to get money out of the model makes one a non-Keynesian, then Joan Robinson was as non-Keynesian as any New Keynesian:

When you are thinking about output as a whole, relative prices come out in the wash – including the relative price of money and labour. The price level comes into the argument, but it comes in as a complication, not as the main point. If you have had some practice on Ricardo’s bicycle you do not need to stop and ask yourself what to do in a case like that, you just do it. You assume away the complication till you have got the main problem worked out. So Keynes began by getting money prices out of the way. Marshall’s cup of tea dissolved into thin air. But if you cannot use money, what unit of value do you take? A man hour of labour time. It is the most handy and sensible measure of value, so naturally you take it. You do not have to prove anything, you just do it.

https://www.jacobinmag.com/2011/07/joan-robinsons-open-letter-from-a-keynesian-to-a-marxist-2/

[Steve Roth]

I think my previous might have been spammified.

@Nick Rowe: "I'm still thinking about this. I'm thinking about red money."

Likewise. In particular I'm thinking about a question that's been vexing me for some time: is your untapped Visa credit-line "money"? Or does it only become money when you lay down your credit card (instead of your debit card)? Any help appreciated.

[Odie]

"If the current period is very short, the consumption function will be almost horizontal (the marginal propensity to consume "b" will be almost zero)."

I would suggest that b is fluctuating between 0 and 1 in very short time periods. Think of a car that travels 50 km from A to B in 1 hour. The average speed is 50 km/h. Now, when you look at very short time periods the speed of the car can be anything between 0 or 100 km/h. You are looking at a distribution around the mean speed that gets more defined the longer time periods you look at. Same with b and the marginal propensity to consume.

"Now let us introduce a stock of money into the model."

Did you not just introduce "a" again? What is money other than future consumption? (And should it not be 0 when looking at the end of human economic activity and assuming rational agents?)

[Nick Rowe]

Odie: when there is a shortage of metal, people switch from metal tools to stone tools, even when metal would be better, and total production drops as a consequence. When there is a shortage of money, people switch from money to non-money, even when money would be better, and total production drops as a consequence.

JW: "But I don't understand how you can be so insistent on this, when right here in this post you have described variation in AD that has nothing to do with excess supply or demand for the medium of exchange."

The model I have sketched here makes no sense except as a model of monetary exchange. If we allowed (costless) barter, unemployed hairdressers would always barter haircuts to get to full employment. The model with a continuum of equilibria was implicitly a red-green money world with a perfectly nominal income-elastic supply (or demand) of money. AD(Y)=Y for all levels of Y, because Md=MS for all levels of Y.

Looking forward to your post. Thanks for all the really great comments here (and on other posts).

[Nick Rowe]

Kevin: JR has forgotten about money as medium of exchange. Yes, I think she is non-Keynesian. I was going to say, in my OP, that NK macro was the bastard offspring of a threesome between Joan Robinson, Knut Wicksell, and Milton Friedman, but then this is a family blog. I was thinking of JR in terms of imperfect competition, but your quote from her there adds more.

[Odie]

"Introduce a fixed stock of land into the model."

Is there a fixed stock of bonds?

"Either the price of land is perfectly flexible or it is not perfectly flexible. If the price of land is perfectly flexible, the price of land jumps up instantly, to eliminate the excess demand for land."

Don't bond prices have a cap so the price is not perfectly flexible?

"If the price of land is not perfectly flexible, there is an excess demand for land, but people will not be able to buy land, because there aren't enough willing sellers. Unable to buy land, people buy consumption instead."

Why can they not just hold money instead and wait for future buying opportunities?

"There are two ways an individual can satisfy his excess demand for money: sell more goods; buy fewer goods. Nobody can stop him buying fewer goods."

Does that not assume the person with the demand for money already has it?

[Nick Rowe]

Odie:
"Is there a fixed stock of bonds?"
No. But it doesn't affect my argument.

"Don't bond prices have a cap so the price is not perfectly flexible?"
No. But it wouldn't affect my argument if they did.

"Why can they not just hold money instead and wait for future buying opportunities?"

If they do, that will cause a recession. Anything that causes an excess demand for money causes a recession.

"Does that not assume the person with the demand for money already has it?"
Yes. But if he didn't have any money he would not be buying anything anyhow, so it doesn't matter.

[Steve Roth]

@Nick Rowe: "Nick E: why stop at M+B? Why not add land in too? M+B+L. Why not add old houses, old cars, paintings, you name it? Where do you draw the line?"

Interesting (to me) that this suggests exactly what I'm suggesting: *all* financial assets (in which I include currency and deeds*) are embodiments of money. The money stock is the aggregate market value of all financial assets.

Which, by the way, is leading me to a rather straightforward monetarist way of thinking about the economy, in terms of the money stock and velocity -- but with a coherent definition of "money."

Changes in desired portfolio allocations result in the total stock increasing/decreasing, but that has nothing to do with "demand for money." All those assets embody money.

* Tricky area, art and other such collectibles as financial assets: can we call your ownership of a Vermeer -- which will never be consumed -- a financial asset? I think it's useful to do so.

[JW Mason]

The model I have sketched here makes no sense except as a model of monetary exchange.

It's one thing to say that there is no such thing as AD in a world without money. It's a different thing to say that AD constitutes excess supply or demand for money.

[Nick Rowe]

Steve Roth: is human capital (labour) also money? Are all newly-produced goods and services, and the ability to produce them, money? What is *not* money, in your view?

[Odie]

>"Don't bond prices have a cap so the price is not perfectly flexible?"
No. But it wouldn't affect my argument if they did.<

Have you ever looked how bonds are priced? You really think I could sell someone a 1-year bond of $1000 at 1% interest for $1011? Not if that person is a rational agent.

>"Does that not assume the person with the demand for money already has it?"
Yes. But if he didn't have any money he would not be buying anything anyhow, so it doesn't matter.<

He could take out a loan in expectation of future income. The problem with demand for money is that some parties have a legal obligation to acquire money while others have a desire for money due to future uncertainty. The two are often not the same.

[Nick Rowe]

Odie: "Have you ever looked how bonds are priced?"

Yes, I have looked at how bonds are priced. Have you ever looked at a first year economics textbook to see the difference between a demand curve that is perfectly elastic at some price, and something that prevents the price rising so that there is excess demand at some price?

"He could take out a loan in expectation of future income."

Taking out a loan means selling an IOU for money.

Just stop.

[Steve Roth]

@Nick Rowe: "Steve Roth: is human capital (labour) also money? Are all newly-produced goods and services, and the ability to produce them, money? What is *not* money, in your view?"

Great question.

First, tangential: human capital is the ability to work/produce. (Stock.) Not synonymous with work/labor. (Flow.)

In my term-of-art definition (in my Celestial Emporium of Benevolent Knowledge), none of those things is money or could even be conceived as money, for multiple reasons. They're all real assets.

Key conceptual distinction in this Emporium, between money and its embodiments:

Money is exchange value embodied in financial assets -- claims on real assets. The assets aren't money (even currency isn't); they're embodiments of money. Money doesn't exist absent its embodiment in financial assets (claims, credits).

Financial assets are distinguished from real assets by the fact that they require no inputs to production and cannot be consumed.

So while the apple on my counter does embody value, it does not embody money, because that value can be consumed, converted into human utility. Money can't. It can only be exchanged.

I find this whole conceptual construct useful because it lets us think coherently like monetarists, with the money stock consisting of the total market value of financial assets (designated in the unit of account). Portfolio readjustments change that market value, hence the money stock. And yes, that portfolio readjustment is often driven by change in demand for currency-like financial assets -- embodiments of money that are historically, nominally stable, relative to the unit of account. But that's not demand for "money." It's demand for certainty, or stability, or similar.

I'm basically trying to distinguish, conceptually, between money and and currency (and other currency-like financial assets), suggesting they're most usefully understood as not being conceptually comparable. That confution is at the root of much of our confusion, IMNSHO. (You'll remember I wanted Mankiw to start his textbook discussing money and value...)

This thinking is not hermetic; I'm still poking holes in it. But I think it's letting me think more clearly about how economies work.

[Steve Roth]

@Nick: "Taking out a loan means selling an IOU for money."

In my emporium:

There's no "selling" involved. Taking out a loan is creating two financial assets -- an IOU and a bank deposit. Each is an embodiment of money. Both balance sheets grow. "Selling" in this context seems like a serious conceptual stretch, perhaps to the point of breaking. Likewise the obverse: buying money for an IOU?

[Nick Rowe]

Steve: if I lend you money, that does not create money. If you repay your loan to me that does not destroy money.

If a (central or commercial) bank creates money, it creates money. It can buy something with that money it has created (including buying an IOU), or it can simply give that money away. A bank buying a computer creates money. A bank giving money to charity creates money.

"Loans create deposits!!!" is a very misleading MMT slogan. Best avoided. Not all loans create money. Not all creations of money create loans.

But this is now totally off-topic.

[Tom Brown]

Nick, an easy one for you. You write:

"when we talk about AD, we are really talking about the excess demand and excess supply of the medium of exchange."

Would Sumner agree? (I'm trying to get straight in my head your differences on MOA vs MOE)

Also, is MOA the "money" in the concept of "long term neutrality of money?"

[Donald Pretari]

Nick, I'm trying to give you a Get Out of Jail Free card. I read the GT a bit everyday, more or less continuously now, but it's all there in the Introduction. Step back and read it rhetorically and philosophically. Is there a stranger Intro that comes to mind? Godel didn't even feel the need to say that.

[Kevin Donoghue]

Donald Pretari: "it's all there in the Introduction." I wonder what you're getting at. Book I of the GT (Chs 1-3) is the Introduction. There's an awful lot there, so what, in particular, do you have in mind?

[notsneaky]

"he will plan to run down his stock of money over time"

Wait, just to clarify. Is this still in the b=1, one period, case? If so, is there some distinction between "model time" and "historical time" (=one) here? (I might be just getting hung up on the "over time" part and getting rid of that in the sentence might not matter for the point being made)

[Maurice Lechat]

A probably naïve question from someone who took second-year macro many years ago.

I’m thinking about the standard Keynesian-cross analysis where an exogenous increase in spending produces an increase in equilibrium national income through the multiplier. What’s happening on the monetary side here? I remember the identity PY = VM. We’re assuming P is fixed, and Y is increasing in this case. Did Keynes think V would automatically adjust for some reason? Or did he assume the private sector held a pool of idle money that would be re-injected into the circular flow of money to make all the additional buying-and-selling at the new equilibrium possible?

And if it’s V that’s supposed to be adjusting, my question is “How could he assume this?” In setting up the IS-LM model there’s this whole thing about V being either determined by the interest rate or determined exogenously by economic institutions. If V is determined exogenously then (if I recall correctly) national income gets determined by the LM curve only and all the Keynesian multiplier stuff becomes irrelevant. Why wouldn’t this thinking also apply in the Keynesian-cross case?

(I think this might be related to some of the discussion above, although I'm not sure.)

[Nick Edmonds]

Nick,

I've greatly enjoyed this exchange. I don't think you've convinced me of your position, but you have certainly made me think more deeply about my own and probably caused me to revise my perspective somewhat.

[Nick Rowe]

notsneaky: "(I might be just getting hung up on the "over time" part and getting rid of that in the sentence might not matter for the point being made)"

Hmmm. I *think* that's right. I think I should have just deleted the "over time" part. It doesn't quite make sense in the context, as you say, but I don't think it's needed to make my point. If you have $100 you plan to get rid of, your permanent consumption will rise by $100.r above your permanent income for the individual experiment. And when you recognise that everyone else is planning to do the same, your permanent income rises by that times the (infinite) multiplier for the aggregate experiment.

Maurice: think of the ISLM model. If V really is fixed (and if M is fixed) the LM curve is vertical, so shifting the IS curve right won't increase AD. Keynes assumed that V was an increasing function of the rate of interest, so the LM curve slopes up, so shifts in the IS curve increase both r and AD.

Nick E: thanks for saying that. I would say exactly the same thing! The discussion here has been very good.

[Nick Rowe]

Don: OK, I have re-read the first 3 chapters to the GT, and I still don't see what you are getting at. My only thought was: "God, this would have been so much clearer if he had simply drawn a labour market diagram, W/P on the vertical axis, L on the horizontal, downward-sloping W/P=MPL curve for the classical labour demand curve, upward-sloping W/P=MRS curve for the classical labour supply curve, where they cross at L*, which can produce a level of output Y*=F(L*), and then said: "OK guys, now what happens if Yd < Y*, so firms can't actually sell Y*?" (It's a shame he never read Patinkin!) And then said that Yd depends on Y, and asked what else Yd might depend on, and whether there would be any process that would bring Yd automatically to Y*?

[Nick Rowe]

Kevin: "What I don't think you're entitled to say is that "New Keynesians are not Keynesians" when, in effect, all they are doing is making money endogenous. If the central bank is following a Taylor rule, M adjusts passively. Since it doesn't really add anything to the model we may as well leave it out."

Hmm. OK. I think I see your point. JMK did in fact say that Yd was a negative function of r, so that the central bank setting the right level of r would get you to full employment. But if the NKs then change Keynes' model, so that it is not the *level* but the *growth rate* of Yd that depends on r, and if that growth rate depends *positively* on r, for a given time path of Y, and if the mpc out of permanent Y equals 1, you would think they might go back and re-ask Keynes' Big Question: "what, if anything, ensures that permanent Yd equals permanent full-employment Y?". And not make that appallingly un-Keynesian intertemporal fallacy of composition.

[Maurice Lechat]

"Keynes assumed that V was an increasing function of the rate of interest, so the LM curve slopes up, so shifts in the IS curve increase both r and AD."

That make sense. So Keynes' baseline model was really the standard IS-LM model. And the Keynesian-cross multiplier analysis is really a kind of warm-up for the real thing, which takes the LM curve into account. But then why bother with the Keynesian-cross multiplier stuff at all? I seem to recall the evidence says the connection between r and V is very weak, which means the fixed-V/vertical-LM story is close to being true. Why not just start there? Wouldn't that make macro a lot less complicated and possibly a lot less controversial?

[Nick Rowe]

Maurice: "So Keynes' baseline model was really the standard IS-LM model."

Well, that is a rather controversial statement, that some "fundamentalist" Keynesians would strongly dispute, but I'm staying out of that one.

You can view the Keynesian Cross model as just a warm-up exercise, as one way to derive the IS curve, when you add an effect of r shifting the AE curve up or down. (The other way is to derive it from the I(r)=S(r,Y) condition.)

I read the evidence differently from you. For example, hyperinflations (when the opportunity cost of holding money is very high) are associated with very high V. Plus, the slope of the LM curve depends not only on the interest-elasticity of the demand for money, but also on the interest-elasticity of the supply of money. And central banks can make that elasticity very high (or very low, or negative), if they want to.

The best case for a vertical LM curve is to assume it is vertical because the central bank makes it vertical, by adjusting the supply of money to keep Y at "potential", to keep inflation on target.

[JW Mason]

Keynes' baseline model has a clear, one-way causal structure:

1. Nominal wages set the price level.

2. Given the price level, the money supply and the (exogenously fixed) stock of "bonds" determine the interest rate. Demand for money varies only due to liquidity preference.

3. The interest rate plus the (exogenously fixed) expectations of entrepreneurs about the income streams expected from newly produced capital assets, determine the level of investment.

4. The level of investment determines the level of output.

There is some contemplation of feedback from 4 to 3, as expectations adjust in light of current output. But variation in output and investment have no effect on asset prices/the interest rate, or on the price level. You could say it's ISLM but with the LM curve horizontal everywhere.

There is a good discussion in chapter 4 of Lance Taylor's Reconstructing Macroeconomics, but in general, this is not controversial. Needless to say, the fact that Keynes said this does not necessarily mean that is logically consistent, or that it is a useful way of describing the economy either in his time or in ours.

[JW Mason]

In other words: If you think money holdings -- and by extension balance sheet positions in general -- are basically about transactions requirements, you draw a vertical LM curve and you're a monetarist. If you think money holdings -- and balance sheet positions in general -- are basically about liquidity, you draw a horizontal LM curve and you're a Keynesian.

[Nick Rowe]

JW: I think I would roughly agree. Keynes seems to "forget" the feedback effect from Y to changes in the demand for money, to interest rates. Even though Y affects the transactions motive for holding money. ISLM closes that feedback loop.

[JW Mason]

Nick-

Yes, that's part of it.

In fact, Keynes leaves out two feedbacks. The one you mention, from Y to asset markets via transactions demand for money. And also, from investment to asset markets, as the creation and financing of new capital assets changes the stock of assets available to wealth holders. Personally, I'm not worried about the first one -- I'm happy to assume that velocity adjusts freely as needed. But the second omission is a problem. There are more or less convincing stories you could tell for why investment decisions don't affect equilibrium in the bond market. But the GT does not provide one.

[Maurice Lechat]

"Plus, the slope of the LM curve depends not only on the interest-elasticity of the demand for money, but also on the interest-elasticity of the supply of money.And central banks can make that elasticity very high (or very low, or negative), if they want to."

OK. But then let me go back to my original question, somewhat revised. If Keynes thought more government spending would help countries get out of the Great Depression was it because:

(1) he thought resulting increases in r would cause V to rise by a lot (in which case, he evidence suggests he was probably wrong),

(2) he thought governments would increase M to prevent the increases in r that would otherwise occur (if so why didn't he just argue for more M?), or

(3) he thought an increase in G would necessarily involve the government drawing from pools of idle money (which the private sector held because rates-of-return were very low and good investments very scarce) and putting this money back into circulation?

[Nick Rowe]

JW: OK. But ISLM also omits that second feedback (in the simple textbook version). It assumes the length of the period is short, relative to the time it takes for the capital stock to adjust (and it does ignore feedback from expectations of future capital stock adjustment).

Maurice: there isn't a lot of talk about fiscal policy in the General Theory. Keynes' case for fiscal policy is more based on your 1, the liquidity trap (which is the same as your 3).

JW could probably give a better answer than me.

[Kevin Donoghue]

Re IS-LM, velocity and all that, I suggest a look at Keynes's GT Ch 15, Section II:

http://www.marxists.org/reference/subject/economics/keynes/general-theory/ch15.htm

[Kevin Donoghue]

Nick: "But if the NKs then change Keynes' model, so that it is not the *level* but the *growth rate* of Yd that depends on r...."

I don't think this is an accurate characterisation of what the NKs do. I've never tried pasting an image in comments here before, but if this works I might try it again. What you will hopefully see is a "simple" version of a NK consumption function, which is not to be confused with the Euler equation. It's the steady-state case with log utility.

[Kevin Donoghue]

Hopefully that monstrosity hasn't done anything harmful to the blog. Obviously Y(t) is disposable income. What I'm getting at is that this consumption function wouldn't have looked so very strange to an Old Keynesian like Modigliani, except for the fact that the sum is infinite. He wasn't too struck on the idea that we should think of households as being immortal and neither am I. In other respects though, it's not much different from his own theory.

[Olli Ranta]

Sorry to be late to the party. As endogenous money was mentioned above I'd like to point that the normal MMT explanations are deficient but it can be done better. Double entry accounting properties causes that some temporary amount will be created when extending loans. That practice then opens the road for substantial amounts. For funding investments we have retained earnings + exogenous money from CB + endogenous from banks. For details see http://olliranta.wordpress.com/endogenousmoney/

[Steve Roth]

@Nick Rowe:

Yes far off topic my apologies. I go here because I'm having trouble buying into the basic terms of the conversation. Or think other terms are more illuminative.

Yes lending by commercial and central banks. (I find it useful to conceive of comms as licensees, subsidiaries of the fed, given how reserve resolution works.)

Those loans -- mortgages, consumer credit, student loans -- dominate the lending in the real economy. $12tn HH debt outstanding compared to $8tn corp bonds. (Not sure about net/gross flows.)

If you buy a $10 watermelon with your credit card instead of your debit card, two financial assets are created -- $10 in bank deposits and a $10 IOU -- that would not have been created if you'd used your debit card.*

You've created and given away a new claim against your real asset, your ability to work, and held on to $10. Those financial assets embody money. You've caused the money stock to increase.

When you pay down that debt, you cause the money stock (total value of financial assets) to decrease.

I think equating the money stock with *currency-like* financial assets is a core conceptual problem, rooted in the conventional-wisdom fairy tale about how "money" was created (coins rather than credit tallies).

Still thinking very hard, can't quite figure, to what extent your unused credit limit can be conceived of as part of the money stock Seems right in the thick of your red-money thinking. If all credit limits were cut in half tomorrow, there would be less spending on real goods...

* The retailer gets the $10 deposit (and gives you the watermelon), but they would have gotten it if you'd used your debit card. So that asset appears on your books (your bank balance doesn't decline), not the retailer's, relative to the counterfactual. Your balance sheet expands, not the retailer's.

[Maurice Lechat]

"Keynes' case for fiscal policy is more based on your 1, the liquidity trap (which is the same as your 3)."

To tell the truth, (3) is only way the Keynesian-cross analysis ever made sense to me. My background is math and stats, so let me use a bit of algebra to do a riff on the haircut-economy model you had in a previous post. Two identical cohorts. Each period one cohort sells haircuts to the other for cash. Only form of saving: hoarding cash. In any period spending by the cohort buying haircuts [Yd(t)] depends on the amount of cash it earned in the previous period [Y(t)]:

(1) Yd(t) = A + b*Y(t) 0 < b < 1.

But Y(t) is just the amount of cash spent on haircuts in the previous period:

(2) Y(t) = Yd(t-1).

From (1) and (2):

(3) Yd(t) - Yd(t-1) = A - (1-b)Yd(t-1).

The system is governed by (3) in or out of equilibrium. The l.h.s. of (3) represents the change in the amount spent on haircuts from one period to the next. Alternatively, this is the amount by which people draw down their reserves of cash to pay for the additional haircuts they want to buy beyond what their current earnings allow. We have equilibrium when the amount of cash exchanged remains the same from one period to the next. This occurs at some Y^ where:

(4) 0 = A - (1-b)*Y^

That is:

(5) Y^ = A/(1-b).

The famous 45-degree line in [Yd, Y] space is the set of all points where Yd(t) = Y(t) = Yd(t-1). At the particular point (Y^ ,Y^) that also satisfies (3) the system reaches equilibrium. My set-up implicitly sets V = 1, because cash-in-circulation changes hands only once a period. The multiplier that comes from (5)

(6) dY^/dA = 1/(1-b)

thus represents both the change in equilibrium income associated with a unit change in A and the corresponding change in the equilibrium amount of cash-in-circulation.