My answers to Steven Landsburg's two questions

Steven Landsburg asks two questions, in two related blog posts. I'm going to give my answers, taking the two questions in reverse order.

Question 2 is here. (It's best you read it direct from Steven.)

Here's my answer:

In microeconomics, if there is a gap between private and social cost, or between private and social benefit (as when there's an externality, or a tax) that normally creates a deadweight cost triangle. But there is one exception, where there is no deadweight cost triangle. That's when the supply curve or the demand curve of the good is perfectly inelastic.

Steven (implicitly) assumes that the nominal stock of money is fixed. Let's keep that assumption. The nominal supply of money is perfectly inelastic with respect to everything, by assumption. But what people care about is not the nominal stock of money M, but the real stock of money, M/P.

In Steven's "Keynesian" world, the price level is also fixed, by assumption. That means the real stock of money must also be fixed, since both M and P are fixed. The supply curve of real money is perfectly inelastic. So any gap between private and social costs or benefits of holding money creates no deadweight cost triangle between the demand and supply curves of holding real money balances.

(There will be deadweight cost triangles elsewhere in the economy if an excess demand for money at full employment causes unemployment, but that is another story, ruled out by assumption in Steven's version).

In Steven's "Friedmanite" world, with perfectly flexible prices, the supply of real money is perfectly elastic. A 10% increase in demand for money will cause a 10% fall in the price level, and so a 10% increase in the stock of real money. So there's a deadweight loss triangle.

It's just like micro: once you recognise that the good people want to hold is not nominal money M, but real money M/P; and once you remember that the size of the deadweight cost triangle depends on the elasticity of supply and demand.

Question 1 is here. (Steven assumes the economy is in a lquidity trap, but that assumption is not necessary for my answer; all that matters is that there is unemployment due to a deficiency of aggregate demand).

My answer borrows from Arthur Laffer, who (I think) said (something like): "Some economists say there's no such thing as a free lunch. Nonsense! Of course there are free lunches. Our job as economists is to find those free lunches and make sure they're eaten!"

The patient had been eating three square meals a day. Then he stops eating lunch, because he wants to save his money. That's bad news. But we know of a free lunch. That's good news that totally offsets the bad news. We are thrilled! But then we hear that the monetary and/or fiscal policymakers won't give him the free lunch. That's bad news again. We are less than thrilled. We are depressed.

If there is unemployment due to deficient aggregate demand, we can print money for free, and give it to people. Holding more money makes them happier. And spending more money makes them happier too. They are doubly happier. And we keep on doing this until they spend so much that aggregate demand is no longer deficient.

Comments

[Mike Sproul]

"Let’s consider the supply and demand for money in this economy."---S. Landsburg

Stop right there. Landsburg's econ book, like all good econ books, is careful to show how demand curves are derived from preference functions and supply curves are derived from production functions. Scarce inputs are used to produce desired goods. But right off the bat we see a disconnect with money. Money can be computer blips. They are not scarce, they are not 'produced' in the production function sense, and they are not 'consumed' in the consumption function sense. In fact, every price theory book that I remember draws demand and supply curves for apples and oranges, but not for computer blips. The concept of supply and demand works well for apples and oranges, but it is not applicable to money. For that matter, it is also not applicable to stocks and bonds, since the value of those things is determined by their backing, and not on the usual supply and demand principles.

[Determinant]

If there is unemployment due to deficient aggregate demand, we can print money for free, and give it to people. Holding more money makes them happier. And spending more money makes them happier too. They are doubly happier. And we keep on doing this until they spend so much that aggregate demand is no longer deficient.

From your lips to Mark Carney's ears, Nick, and Jim Flaherty's and Stephen Harper's too. Weren't they in the same room together not long ago? Damn it man, why weren't you there ???

Except economists for the last generation and the politicians who listened to them had a counter idea that cranking up the printing press would create this big, bad, nasty boogeyman called Inflation. It would eat your savings alive and do nothing to cure unemployment. Economists made this argument so well that the generation now in power will not, absolutely not pull the Money Lever. They can't stand the thought of the Boogeyman called Inflation.

We keep telling them that the boogeyman doesn't exist but they won't listen.

[Scott Sumner]

Great post. A while back I published a paper that sort of made the same point, but from a different angle. Recall the standard Keynesian view is that as the LM curve gets flatter, the case for monetary stimulus is weakened and the case for fiscal stimulus is strengthened. I argued that this view had things exactly backward. The flatter the LM curve, the more money that is needed to boost AD by X%. But that's good, as monetary stimulus has a negative cost. You are buying interest bearing bonds with non-interest-bearing currency. That's a free lunch. So the flatter the LM curve the weaker the argument for fiscal stimulus, and the stronger the argument for monetary stimulus.

Of course some would object that the central bank might later lose money if it pulled currency out of circulation once the country was no longer in a liquidity trap. But the EMH says they shouldn't be expected to lose money, so I'm not particularly concerned about that risk.

(My comment refers to question 1)

[Too Much Fed]

"If there is unemployment due to deficient aggregate demand, we can print money for free, and give it to people."

Exactly how would that happen?

Are you assuming an aggregate demand shock or an aggregate supply shock?

[Patrick]

On question 1 ... Doesn't Keynes basically answer Landsberg's question right out of the gate in Chapter 2?

[Ralph Musgrave]

Steven Landsburg’s point is relevant to the question as to how far banks should engage in maturity transformation. Banks make maximum use of depositors’ money, because from the private perspective, money is an asset (though, as Steven rightly points out, it’s not an asset for society as a whole). That is, banks compete to borrower ever shorter and lend ever longer, and we all know the end result: Northern Rock, etc. This competition is fatuous from the point of view of society as a whole.

Therefor (so I would argue) maturity transformation should be curtailed. The way I’d like to see it curtailed is to force depositors to choose between two types of accounts. 1, “transaction” or “instant access” accounts. That would be money that depositors clearly have no intention of seeing tied up in investments. That money would be lodged at the central bank, where it would be 100% safe, but would earn little or no interest. 2, “investment accounts”. Here, the money CAN be loaned on, plus it would earn interest, but there’d be no taxpayer funded rescue if it all goes belly up: after all, it’s not the taxpayers’ job to subsidise commerce.

[Nick Rowe]

Mike: furniture, paintings, and houses are not "consumed" either, in that sense. We get benefits from holding those assets though. Just like the stock of real balances we hold. And the amount people want to hold depends on the cost of holding them. The utility of holding money is like the utility of owning shopping bags. They make the shopping easier.

The supply of money is another story. For some weird reason, we have weird ways of supplying money. We used to dig it up out the ground. Now it's mostly produced as some sort of accidental by-product of producing loans.

Determinant: But Carney and Harper do understand all this, and, very roughly, have gotten it right. (Except, they were behind the curve in 2008.) I don't need to tell them this. (I did tell them they were behind the curve in 2008.)

Does an increase in the money supply cause increased employment, or does it cause inflation with no change in employment? Both answers are right at the same time. Lucas showed us how to understand this paradox. Carney (almost certainly) understands this. (Some of his deputy governors certainly understand this.)

Scott: Yep. A small money multiplier increases the total benefits of solving a given-sized deficient AD problem with monetary policy.

TMF: either a fall in AD or a rise in AS would, at the old price level, create a deficiency of AD.

Patrick: I grabbed the GT from my other desk, skimmed Chapter 2, and I would say "no".

[Nick Rowe]

Ralph: I think Steven Landsburg had in mind a simpler model in which M was directly under the control of the government.

[Kevin Donoghue]

Patrick: "On question 1 ... Doesn't Keynes basically answer Landsberg's question right out of the gate in Chapter 2?"

He certainly disposes of the claim that "there are no utility functions in the old Keynesian model" which is a bizarre assertion. But perhaps Landsburg believes that when Keynes rejects the claim that "the utility of the wage when a given volume of labour is employed is equal to the marginal disutility of that amount of employment" he is, in effect, banishing utility functions.

AFAIAC it's up to Landsburg to make that case.

[Kevin Donoghue]

I had the feeling that Steven Landsburg was simply confusing himself and to his credit he has acknowledged that; case closed, I think:

http://www.thebigquestions.com/2011/09/28/doh/

[Luis Enrique]

Shouldn't he be comparing the private cost of holding money with the social cost of holding money? Is the social cost of holding money the same as the social cost of supplying it?

I don't understand what these dead-weight cost triangles would be doing in any case. He says spending too much money bids prices up. What do we care what the price level is? Where is the real effect?

I confess I don't understand the whole thing. He calls the cost of holding money forgone consumption. But haven't I optimised my consumption? So what am I forgoing? I'm holding money because doing so has value to me. Haven't I chosen a point at which private marginal benefit equals private marginal cost? If you somehow removed the private cost of holding money, so I could move my optimal choice, would you not also be removing the value to me of holding money? Still don't see how we have a suboptimal outcome in this Friedman world, but supposing we do, is it only suboptimal in comparison to a world that cannot exist?

[Mike Sproul]

Nick:

"The supply of money is another story. For some weird reason, we have weird ways of supplying money. "

Amen. Trillions of units of computer blip money can be issued in one moment and retired in the next. With paper money, the same process takes minutes instead of nanoseconds, but that's still plenty fast. This makes it meaningless to speak of "money supply". But no matter how fast money is issued and retired, every unit of money is the liability of its issuer, and the retirement of the money extinguishes that liability.

[Ralph Musgrave]

I agree with Luis Enrique, in particular when he says “I'm holding money because doing so has value to me. Haven't I chosen a point at which private marginal benefit equals private marginal cost?”

Put another way, there are private net BENEFITS (not costs) from holding money in as far as that stock of money satisfies the well knows precautionary and transaction motives for holding money. In contrast, if the stock of money rises above that level, THEN AND ONLY THEN do private costs begin to exceed the benefits, and the private sector will try to dispose of its surplus stock of money.

[Nick Rowe]

Kevin: good find. On a quick read though, I'm not sure I understand or like his answers.

Luis: here's one way to think about it. Real balances are an interest free loan to the government supplier of money (assume all money is zero interest currency). If i is the rate of interest on government bonds, and there's zero cost of printing new bills to replace worn ones, etc., then it is as if the government were taxing me i(M/P) per year. The deadweight cost triangle is the real resource cost of my attempts to economise on holding money, because I equalise my marginal private benefits to my marginal private costs. "Shoe-leather costs" is how that triangle is normally caricatured. I run back and forth between shops and banks more frequently in order to reduce the size of the interest-free loan I am giving the government.

[Luis Enrique]

Thanks for the reply NIck,

well, if I was exerting effort to minimise the money I hold in my wallet, that would certainly count as a real cost that could be eliminated by creating an interest-bearing means of payment for me to use at my leisure. However Steven Landsburg says the problem with people spending too much money is that it bids up prices. I don't see how to square that with your explanation. Nor yet understand what Landsburg thinks the real cost of a higher price level is.

a small point: is the opportunity cost of holding money (not holding bonds) really the same thing as making an interest-free loan to the government?

[Kevin Donoghue]

Nick: "I'm not sure I understand or like his answers."

Do you mean Landsburg's, in his "D'oh!" post, or Keynes's in GT Ch2?

Now that Landsburg has an answer he's happy with, I think I see what his question was; and all credit to Jim Kahn “who instantly saw the answer so clearly that he couldn’t figure out what the question was, and expressed it in language that forced [SL] to see the light.”

As best I can tell it's as simple as this: start with a simple one-period model so that we don’t need to worry about interest rates and such. In a Walrasian equilibrium real balances M/P are always optimal. It doesn't matter how thrifty people are, the auctioneer can always get to that optimal value through deflation. But if prices are sticky, the only way to get the optimal M/P level (and hence full employment) is to print money.

That's Keynes's answer too, except that (1) he writes Mw = M/W and calls it the money supply expressed in wage-units; (2) he calls a Walrasian equilibrium a long-run-we're-all-dead equilibrium; and (3) he hammers the point home with a memorable line about giving people green cheese when they want the moon.

[Ralph Musgrave]

Luis Enrique is right to question whether money is a loan to government. Obviously money (or to be accurate, monetary base) is NOMINALLY a loan to government, that is, monetary base appears on the liability side of central banks’ balance sheets. But this is an entirely ficticious liability. As Willem Buiter put it, “These monetary (base money) ‘liabilities’ of the central bank are not in any meaningful sense liabilities, because they are irredeemable.”

[Nick Rowe]

Luis: "well, if I was exerting effort to minimise the money I hold in my wallet, that would certainly count as a real cost that could be eliminated by creating an interest-bearing means of payment for me to use at my leisure."

There are two ways to eliminate that deadweight cost triangle: one is to pay interest on money equal to the interest on bonds; the second is to create sufficient deflation to bring the equilibrium ("full employment") nominal interest rate on bonds down to 0%. This is all based on Milton Friedman's "Optimal Quantity of Money" argument (which is quite different from what he proposed in more practical policy advice; it was more of a theoretical thought-experiment.)

"However Steven Landsburg says the problem with people spending too much money is that it bids up prices. I don't see how to square that with your explanation."

Start in equilibrium, where individuals equalise the marginal private benefits and costs of holding money. Then suddenly they all get altruistic and decide it's OK to give the government an interest-free loan. They all hold more money, so P falls until M/P is higher. There is now less waste of shoe leather. Steve is just going in the opposite direction, starting in the altruistic equilibrium, then moving to the selfish equilibrium.

"a small point: is the opportunity cost of holding money (not holding bonds) really the same thing as making an interest-free loan to the government?".

Yes, for the individual. If I hold $100 more currency and $100 less bonds, which pay 5% interest, I lose $5 interest and the government pays $5 less interest. Plus my I save $5 on shoes, at the margin.

Kevin: I meant Lansburg's "D'oh".

[Min]

Nick Rowe: "If there is unemployment due to deficient aggregate demand, we can print money for free, and give it to people."

Too Much Fed: "Exactly how would that happen?"

In many ways, eh? How's this? Every morning at every homeless shelter give everyone who stayed overnight $100. :) That's probably not enough, but it's one way.

[Min]

Ralph Musgrave: "money is an asset (though, as Steven rightly points out, it’s not an asset for society as a whole)."

Please explain. OC, it is not an asset in the sense that society as a whole buys anything with it, but isn't useful?

[Ralph Musgrave]

Min, Yes you are right: money is an asset in the sense that most economies are better off with a form of money than using barter. But the point I was trying to make above was that dollar bills are inherently worthless. Or put another way, a country where everyone wants to hold $1,000 for transaction and precautionary motives (and actually has that $1,000 per person) is not for that reason better off in real terms than a country where everyone is happy holding $500.

[Luis Enrique]

Nick, thanks again for the answer. I'm getting there! Although I still think what you have described holding money as is not-lending the government money, not lending them money.

[J.V. Dubois]

Excellent article and very good discussion (as usual). However I still have one question as a followup of shoe-costs and free loan to government. Given our current system, when people mostly only hold money on bank accounts can it be counted as a loan to the government? With fractional-reserves system people basically hold a bank (not government) liability which can be (theoretically) redeemed using ATM.

So as per some older post, if I start issuing personal JV dollars (or loans in those dollars) and I promise to convert them at par to regular government dollars, is it not me who now reaps the benefits of the money as costless funding source? If I could get the government to establish an institution serving as a lender of last resort who would back this redeemability, I would basically become a monopoly provider in this newly privatized market for cash, effortlessly gathering the rent created for me by the government.

[Min]

Ralph Musgrave: "the point I was trying to make above was that dollar bills are inherently worthless."

OIC, Thanks, Ralph. :)