Is Spending Money a Market Failure?
Steve Landsburg is trying to think through the implications of assuming sticky prices. But it’s the first part of his post that strikes me as fishy:
[L]et me start with something I do understand — a world with flexible prices and (for simplicity) no inflation. (What follows is standard textbook material, which we all learned from Milton Friedman, who in turn probably learned it from Iriving Fisher. It is not controversial. If you think it’s wrong, the probability is 99.9999% that you’re mistaken.)
Let’s consider the supply and demand for money in this economy. Money is supplied at (essentially) zero social cost (the cost of paper and ink being essentially zero). However, the private cost of holding money — measured in forgone consumption — is positive. Whenever the private cost of an activity is greater than the social cost, people engage in too little of that activity. In this case, they hold too little money. Or in other words, they spend too much money. That means that each additional dollar you spend must hurt your neighbors more than it helps them. It remains to ask who, exactly, is hurt by your spending.
Answer: When you spend a dollar, you bid up prices. That’s good for sellers, bad for buyers, and bad for other people holding money (because it depletes the value of their dollars). The first two effects wash out, because every transaction involves both a seller and a buyer, leaving the moneyholders as the bearers of the net harm.
Note the logic: First we identify a discrepancy between private and social cost. This tells us that spending a dollar must (at the margin) do more external harm than good (”external” means “not felt by the decisionmaker”, the decisionmaker in this case being the spender). This in turn tells us that we ought to ask who bears that harm. The answer isn’t immediately obvious, but it’s clear once it’s pointed out.
I’ve often thought that I’m one-in-a-million, so I take up Steve’s gauntlet while acknowledging the intrepidity of assaulting Irving Fisher. (Then again, he was wrong on Prohibition and the 1929 stock market.)
At first I tried to apply Steve’s logic to coat-check tickets. (In other words, would Steve say that the private cost of holding–rather than cashing in–a ticket to get your coat back, is higher than the social cost?) It wasn’t clear how to make the analogy work, since the issuer of the tickets can’t just print them up at zero cost; they are supposed to be backed by a coat.
So this led me to think about commodity money. Would Friedman’s analysis apply to an economy with gold as the money? (If so, note that in a Walrasian model you can pick any good as the numeraire.) I guess not, since a major difference is that you have to mine the gold.
OK so it’s clear that for this analysis to “work,” we have to be talking about fiat money. But then that led me to a new objection. For the sake of brevity, let me just ask: Steve, what’s to stop me from writing the following?
Let’s consider the supply and demand for money in this economy. Additional money is supplied that confers (essentially) zero social benefits (since having more pieces of paper in circulation doesn’t make the community actually wealthier). However, the private benefit of holding money — measured in liquidity services — is positive. Whenever the private benefit of an activity is greater than the social benefit, people engage in too much of that activity. In this case, they hold too much money. Or in other words, they spend too little money. That means that each additional dollar you spend must help your neighbors more than it hurts them. It remains to ask who, exactly, is helped by your spending.
Answer: When you spend a dollar, you bid up prices. That’s good for sellers, bad for buyers, and good for other people holding goods and services (because it increases the market value of their goods and services). The first two effects wash out, because every transaction involves both a seller and a buyer, leaving the goods-holders as the bearers of the net benefit.
Note the logic: First we identify a discrepancy between private and social benefit. This tells us that spending a dollar must (at the margin) do more external good than harm (”external” means “not felt by the decisionmaker”, the decisionmaker in this case being the seller). This in turn tells us that we ought to ask who bears that benefit. The answer isn’t immediately obvious, but it’s clear once it’s pointed out.
If Steve doesn’t like that angle, let’s riff on Friedman’s permanent income hypothesis, applied to cash balances:
OK let’s stipulate for the sake of argument that when people spend money, they are imposing negative externalities on people. But by the same token, when people acquire money, they must be imposing positive externalities on people. Since no one (outside of Bernanke) is a counterfeiter, these effects cancel out over the course of each person’s lifetime, properly adjusting for interest rates, passing on property to heirs, etc.
Last point: I am not bothering to look up Friedman’s classic essay on this issue. I am fully prepared to admit that I am wrong, within the context of a standard neoclassical model. But Steve’s post isn’t making me “see” it…
When you spend money on a thing you increase demand for that product which causes more of it to be produced by that company and its competitors which, due to the beneficent effects of competition, ultimately results in lower prices for said product and therefore an increase in the standard of living of all that use that product.
Isn’t this Econ 101?
Apparently I don’t know how to spell my own name.
Bob: I think it’s clearer if you think of the market failure not as an externality, but as the deadweight loss due to a tax (or if you prefer, to a monopoly price).
The govt prints up money at zero marginal cost, and sells it to you at a positive price (they’ll only give you a dollar if you give them an apple in return).
Therefore you “buy” too little money. It’s exactly as if the government could produce, at zero marginal cost, a miraculous wrinkle cream for which they charge $5 an ounce. You’d buy too little. This means that if, after you’d reached your optimum, you were forced to buy more, total surplus (i.e. govt revenue plus your consumer surplus) would increase.
As for your clever alternative accounting: You are imagining an increase in the money supply and then, in order to conclude that there’s no social benefit, you’re letting the price level adjust. But the price level adjusts precisely because people are refusing to hold this money. So your thought experiment is analogous to one where people refuse to buy extra wrinkle cream. But to prove that the actual outcome is suboptimal, we want to compare it to the counterfactual case where they (foolishly) do buy more wrinkle cream.
The right thought experiment, then, is to ask what happens if the authorities supply additional dollars and people hold them. In that case, the social cost of the additional dollars is zero; the private cost (forgone consumption or forgone interest) is positive, and the private/social benefit is the increase in real balances. Since SB > SC, this is a good thing, but since PB is not > PC, people won’t go for it. That’s why the good thing doesn’t happen. Instead people refuse to hold the dollars, dissipating the potential social benefit.
In other words, you are accounting for what does happen; I am accounting for what would happen if you could somehow get people to hold these dollars. My calculus shows that in that case, things would be better.
OK Steve I understand what you are saying; I need to let it sit for a while.
What does he man by ‘if you think it’s wrong’? Obviously prices aren’t flexible and inflation isn’t 0?
I guess most neoclassical economists are beyond help.
I’m having a little trouble with the allocation of costs here. First, printing of money isn’t “zero social cost” as I usually think of social cost. It’s zero private costs to the monetary authority. So thinking about this as a private vs. social cost accounting seems wrong. But then with the money holder it’s not clear that a positive cost associated with staying liquid means they’ll engage in too little of that activity because that is balanced on the margin with the benefit of staying liquid (namely, having a cushion for uncertain times). That private benefit isn’t discussed here, and it seems like a lot of this “too much for optimality” or “too little for optimality” talk relies on one private cost and zero social cost. This seems to essentially be your thought too.
And of course if we think of neoclassicism a little more broadly (ie – beyond Friedman and Fisher… although Fisher was close to getting there in the 1890s) we can also think about additional compensation for parting with liquidity but not consuming – namely, earning interest. For obvious reasons I think that’s where somewhat more of the action is. The benefits of staying liquid increase, so the compensation for parting with liquidity increase, and that influences a lot of other decisions by investors.
Daniel Kuehn: Yes, interest is definitely where the action is. The argument assumes a positive nominal interest rate. If the nominal interest rate is zero, then the private cost of holding money becomes zero, bringing it in line with the social cost, so people hold the optimal amount.
That’s why Friedman argued that the optimal inflation rate is equal to minus the real rate of interest —- to make the nominal interest rate zero.
So I’m still not clear on why you’re calling the production cost a “social cost”. Could you explain that?
I could imagine some relevant social costs and benefits here. Dani Rodrik has talked about exchange itself as a social benefit (i.e. – a network externalities type argument), and that seems relevant here as a social cost of staying liquid. But I don’t understanding why you’re calling private production costs (even though they’re done by the government) the “social cost”.
You also seem to be not considering the private benefit of staying liquid (which is odd, since you were riffing on that recently in another post!). Let’s say the private benefit is quite high due to, say, uncertainty about the future path of prices and demand. And let’s say there’s a social benefit to exchange (Rodrik’s point). Now we’ve got a positive externality and too LITTLE spending.
I feel like your discussion is considering all the right margins, but your conclusions are extremely sensitive to what costs and benefits you’re considering private vs. social, and which costs and benefits you’re choosing to consider.
Daniel: Of course private production costs are part of social costs. How could they not be?
Steve: If you’re still checking this, can you quickly tell me what you think about viewing things from the flip side? I.e., why isn’t every change in money-prices a perfect wash for society as a whole? You recognize that a price change is a wash for “buyers and sellers,” but then seem to inconsistently just look at money-holders after that. But why not look at goods-holders and say their gains perfectly balance the losses of the money holders?
Also, do you agree with me that if it’s a negative externality to spend a dollar, it must be a positive externality to accept a dollar?
Last thing (if you’re really really bored), can you confirm or deny whether your analysis works if we’re using a commodity money?
Bob: The difference between goods holders and money holders is this: When you change the price level, you don’t change the actual quantity of apples out there, but you *do* change the actual quantity of real balances. So the “gains” to the goodsholders are in that sense nominal, while the gains to the moneyholders are real.
There is definitely a positive externality when you accept a dollar. If I sell you an apple for a dollar and then put that dollar in my mattress, and then the price level adjusts, then the rest of the world (including you) is up one apple, but not down anything, ,because the ROW’s real balances recover their original value as a result of the price level adjustment.
If Scrooge McDuck likes filling his vault with paper money so he can swim in it, and is willing to exchange goods for that paper money, then the rest of the world is essentially getting those goods for free. True, some people briefly sacrifice some real balances to acquire those goods, but as soon as the price level adjusts, the non-Scrooge part of the world has all its real balances restored. Most of that gain goes to people who were never parties to the original transaction, so I think we should certainly call it an externality.
I don’t think any of this works with commodity money.
Last thing: This is not “my analysis”; it’s Milton’s.
Steve (Landsburg): OK one last thing, and then I’ll probably do a follow-up post. I like your Scrooge example. So what’s wrong with saying this?
* When Scrooge accepts a dollar, he helps the rest of the community. If Scrooge never spent that dollar, he would have given a gift to everybody else.
* When Scrooge spends a dollar, he hurts the rest of the community.
* If Scrooge sells a TV for $100, then next week buys two DVDs for $100, on net he has neither harmed nor helped the community. He’s ultimately just exchanged a TV for two DVDs.
* Unless we’re dealing with counterfeiters or people who don’t give their cash balances to their heirs, each person during the course of his life (in PDV terms) neither helps nor harms the rest of the community through his buying and selling of dollars.
I’m not sure why you’re asking this – of course private costs are part of social costs and I don’t think I’ve said otherwise. The problem is you are taking one private cost (production costs), calling that the “social cost”, another private cost (holding money), and not counting that as a social cost, and you’re ignoring many private benefits as well as other social costs and benefits. The way you’re doing the cost accounting here is strange.
Production is not completely costless, but it’s mostly costless and what cost there is is covered be seinorage. Fine. Now we have a bunch of money. What is done with that? Well there are private opportunity costs associated with holding it and there are private benefits associated with holding it. And if other parties want you to part with your liquid holdings they’ll pay you interest to raise the opportunity cost of holding money. Fine, that’s all good.
Everything seems to me to be functioning just fine unless we can think of other social costs or benefits that might make this sub-optimal.
The one that immediately comes to my mind is the network externalities associated with trade.
I can’t think of any non-private (ie – non internalized) social costs that might justify your suggestion that we “spend too much”. The only social cost you mentioned seems to me to be fully internalized.
Daniel: You went wrong when you said “whatever costs there are are covered by the seignorage”. In fact, they are *more* than covered by the seignorage. The seignorage is mostly pure social gain. If people were willing to hold more money, there would be more seignorage and hence more social gain.
OK fine – I don’t know the ins and outs of seingorage. The point is noting one private cost of holding money and one private cost of producing money doesn’t seem suffiicient for determining whether we spend too much.
Do you have any thoughts on the private benefits of holding money? The (externalized) social benefits of spending money? Seignorage doesn’t change my argument about those things.
I share Daniel’s confusion. It is true that production costs of fiat money are low. But what is your social utility function? When the government adds money to someone’s cash balance, they seem to be “costing” the people who did not get the new money. Let’s say Alice and Bob are saving up to buy houses one year from now and they save the same amount each month.
Scenario 1: Bob saves more than Alice. One year from now, he bids a higher price than Alice. He buys the house. Alice is left with a cash balance. The seller makes the purchase price.
Scenario 2: We “help” Alice by boosting her cash balance with fiat money drops. In one year, she will have saved more than Bob. She is able to buy the house since she can offer a higher price. Bob is left with a cash balance (bigger than Alice’s balance in Scenario 1). The seller makes a higher purchase price.
[Alternatively, Alice could hit her “savings target” faster than Bob and simply start bidding sooner than one year from now. Bob will not be in the market yet. Or he can race to market with whatever he has saved so far.]
Either way, doesn’t Bob “pay” for the money drops by not getting the house? Shouldn’t this be in the social utility function?
Marris: People care about two things: their houses and their real balances. In your scenario, the number of houses never changes, so the only changes in social welfare come from changes in real balances. As long as Alice is willing to hoid more money than previously, her real balances are increased and nobody’s are decreased. Hence the increase in social welfare.
What do you mean? “Real cash balance” is nominal cash balance divided by some average transaction price over a _given_ period, right?
Bob’s real cash balance _this period_ is his nominal cash balance this period [or at some point in this period] divided by some average transaction price for this period. Does he really care about that? He cares about his ability to buy the house. At most, he cares about the average transaction price in the period in which he thinks he will spend.
What happened here is that the house was bid away from him in one year because of the money drop effects that took place _in that period_. And he cares about that.
The same is true for every saver. In your example above, where Alice saves money, it’s *possible* that she is planning to save “for the sake of increasing her cash balance and keeping it there forever.” But that’s really weird, right? She is probably saving so that she [or maybe her heirs] can spend it in some future period.
I admit the Scrooge example does not exhibit this. But that’s because Scrooge is saving for “direct consumption” (so he can swim in it) and _while swimming_, he does not care what the money would be buying if he spent it.
Is the discussion about entrepreneurial reallocation of resources for production, or the actions of market participants who are not shifting their demand preferences? My rudimentary understanding is that the equilibrium price is determined by the intersection of the supply and demand curves. If we are talking about prices that stay in equilibrium in the long run, then I don’t see why there would be any externalities associated with purchasing power
Bob: Responding to your comment that starts “Ok, one last thing…”
I’m with you up to the last paragraph. If Scrooge, over the course of his life, keeps 10 million green pieces of paper in a vault, and if he exchanged real goods for those pieces of paper, then he does, over the course of his life, help his neighbors.
It’s good to mention counterfeiters, because everyone can see how counterfeiters do harm. Scrooge is an anti-counterfeiter; instead of adding money to the system, he takes some out. If counterfeiters are bad, Scrooge must be good.
That really is the entire point: To demand money is to be an anti-counterfeiter. You demand money up to the point where it serves your own purposes to do so, but you ignore the social benefits of being an anti-counterfeiter; therefore, you don’t demand enough.
OK I totally get your perspective and I’ll try to do a follow-up post in the near future. But now I have to go do some anti-counterfeiting (so that I can harm everyone next month when I pay my mortgage).
Bob Murphy,
Since you understand Landsburg’s point, would you be kind enough to fill us in? If you want to just write the article, that’s fine. What I don’t understand is:
(1) What properties are in Landsberg’s social utility function?
(2) Is there some faulty treatment of time periods? That is, does Landsburg think “no one is worse off this period” because Alice and Bob have not yet started bidding for the house [see example above]. If we take this approach, it seems that Bob has not lost out this period [he has not lost the chance to buy the house yet] and Alice has gained [more money! yay!].
But what’s really happened is that Bob “does not see” that he’s been made worse off. And he won’t see it until next year. A social utility function which does not evaluate “processes” will not capture this loss.
Marris: If you really want to understand this, you should read Friedman’s “optimum quantity of money” paper.
I’ve only found this book [http://www.amazon.com/Optimum-Quantity-Money-Milton-Friedman/dp/1412804779]. If possible, can you (or someone) post a link to the paper?
I’ve often thought that I’m one-in-a-million
Which would leave us with about 7000 Murphs. That would be cruel. At least your fans are kind enough to think you are unique.
There’s probably a lot to say about this (but I am way too confused) but here are a few quick thoughts:
1) Isn’t there only an external effect in monetary disequilibrium? It’s only when average real balances supplied exceed those desired that “spending” alters the price level. The transactions demand (the average demand to hold money for “spending” purposes) is obviously an important source of demand for money and in fact of the social benefits of money.
2) Money, and especially unbacked paper money, is a network good. Expectations regarding the demand for money by others and the potential for monetary disequilibrium necessarily one’s own demand for money.
3) The thing that is throwing me off is the idea implicit in this discussion that a greater supply of money necessarily provides a greater supply of “money services” and therefore we should try to get people to hold a greater supply of money, quite independently of the need to maintain monetary equilibrium, because of its low production costs. That strikes me as incomplete.
4) The external effects of individuals responding to monetary disequilibrium by trying to get rid of or acquire money balances certainly might be, along with other discoordinating effects, reasons to maintain monetary equilibrium, but I am not sure why that is related to the absolute costs of paper money production (as long as the costs of adding money to maintain equilibrium, in periods of excess demand, exceed the minimal costs of money production).
5) If the nominal interest rate is maintained at zero, and one can, in effect, earn the real rate simply by holding money, what would be the compensation for a contractual shift of liquidity and the attendant loss of flexibility (i.e., lending)?
I’m a little confused by all the big words here, but I think I agree with the writer’s point (if it is his point) that hoarding money (ie high cash demand) is a net benefit to society in our monetary system. Holding cash, instead of depositing it at a bank, should lead to lower inflation because banks can’t expand the money supply at the rate that money was in the form of a demand deposit account. This should lead to deflation of prices., which would ultimately change the equilibrium level of cash demands anyway
That’s from Rothbard’s Mystery of Banking (or my understanding of it anyway). Again, I don’t know if it’s relelvant to this discussion. I guess, in summation, hoarding cash > depositing cash at a bank.