Suppose your country (call it "Greece") is in recession, because there is an excess demand for money (call it "Euros"). And suppose that the Euro is both Medium of Account (prices are quoted in Euros) and Medium of Exchange (all other goods are bought and sold for Euros).
Now suppose your government introduces a new currency (call it "Drachmas"). It introduces the new currency by paying pensions via helicopter. But the Drachmas do not displace the Euros. The Euro remains as medium of account. Prices are sticky in terms of Euros, but the exchange rate between Drachmas and Euros is perfectly flexible, so prices are perfectly flexible in terms of Drachmas. Both Euros and Drachmas are used as media of exchange.
What happens? Does the introduction of the Drachma make the recession less severe?
I say it does make the recession less severe.
There are two ways to explain why it makes the recession less severe:
1. Unless the Drachma has an exchange rate of zero, the introduction of the Drachma, in addition to the existing stock of Euros, increases the total real value of the stock of media of exchange, and so lessens the excess demand for the media of exchange, and so reduce the severity of the recession. And if at least some people are willing to use at least some Drachmas as a medium of exchange, there will be some demand to hold Drachmas, and so the exchange rate of the Drachma will not be zero.
2. If at least some people are willing to use at least some Drachmas as a Medium of exchange, that means that Drachmas are an (imperfect) substitute for Euros. The introduction of a new good will reduce the demand for any existing good that is a substitute for the new good. So the introduction of the Drachma reduces the excess demand for the Euro, and so lessens the recession.
But in order to get this point, you have to understand the essentially monetary nature of recessions. They aren't caused by real interest rates being wrong, or real exchange rates being wrong, or real wages being wrong. Those are all just symptoms, or side effects. The underlying cause of the recession is the reduction in the volume of trade caused by an excess demand for the one good that is demanded or supplied whenever any other good is traded.
If you want a model, use my "Tiny money/macro model for microeconomists", and ask what would happen if we started out in recession, so there was too little trade in apples and carrots, and then introduce an endowment of a fourth good, dates, that was a substitute for the bananas that are used as a medium of exchange. Even if the prices of apples and carrots stayed exactly the same in terms of bananas, and the price of dates were perfectly flexible, the volume of trade in apples and carrots would increase.
This is in part a response to JP Koning's good post. He may be right about the Euro remaining the medium of account, but it is the excess demand for the media of exchange that causes recessions.
Worthwhile Canadian Initiative 
suppose we used cows as money. An increased demand for milk (plus sticky prices) would cause a recession, because people would stop spending their cows because they didn’t want to lose the milk. If the government helicoptered in some dry cows (that gave no milk) that would cure the recession. Dry cows would trade at a discount, but people would spend them.
This again assumes the conclusion that there is an excess demand for cows-to-trade in the first place. What happens when enough dry cows fall and there is no marginal demand for dry cows at all but the demand for milk is still rising? Won’t there still be downward pressure on the price level amid nominal rigidity?
You are adding an unnecessary wrinkle with this story, because the helicoptered dry cows are also, by definition, marginal wealth And there is always a demand for marginal wealth. This is why the cow/drachma analogy doesn’t work: There is no such thing as a completely “dry drachma,” since milk in this story = safe savings, and for the drachma to have any MOE use it must have some safe savings feature. That’s why I asked you to imagine the government successfully announcing that existing shares of stock would heretofore be used as an additional MOE (but not MOA). Otherwise you are conflating the issues. Helicoptered drachma with no marginal MOE demand might work anyway as fiscal stimulus, but that is different than saying you are relieving excess demand for the MOE.
Do you at least agree that you can imagine a world where the cow government successfully converts (jawbones) all existing shares of stock into an additional MOE (but prices are still quoted in cows and cow/stock prices are flexible as usual), people are fully satiated in liquidity, but a spike in milk demand causes a very familiar recession (more sellers than buyers, because, of course, sticky prices are too high relative to the price of cows — not because of a lack of MOE)? Doesn’t this tell you that Drachmas-as-MOE might have limitations as a cure just as more base money very quickly became irrelevant in the US in 2009?
Nick,
You know better, New Keynesian models have nothing to do with traditional Keynesian models.
It seems like a leap from “ISLM model” to “traditional Keynesian model”.
Avon: you know better than to use the patronising “you know better” style of argument. Are you accusing me of lying? Actually, New Keynesian models do have a lot to do with ISLM. Some economists don’t see this, but I do.
john: “Nick, if the greek debt is denominated in euros then wouldn’t the greek state have to pay interest on that debt (and repay the debt eventually) in euros?”
Yes.
“Doesn’t that imply that the greek state would REALLY prefer taxes beeing payed in euros instead of new drachmas?”
No. As long as they have the same value, at the current exchange rate, the state won’t care.
“If that’s the case, and everybody knows that, won’t everybody be hoarding euros in order to lend at the state on the date it needs to make a repayment?”
No more than they currently are hoarding Euros to pay taxes.
Nick Rowe,
As you put it once, New Keynesianism is a conspiracy to convince central bankers to make the RBC/Walrasian/loanable funds theories seem true.
dlr: maybe this old post explains my argument better.
Nick, I remember that post, and I don’t agree. It has some similarities to your even earlier rent-control argument. It misses the MOE vs MOA issue, in my view. In that post, you seem to assume that Barter only means “no medium of exchange,” and so conclude that MOE is the key. But that is not what barter means. Barter means no medium of exchange AND no medium of account. This is required because at the barter limit where every item is fully liquid, it makes no sense to have an MOA. Direct ratio exchanges in barter trumps an MOA. So you eliminate both the MOA and MOE, but somehow conclude that this shows the MOE is what’s important. Imagine an impossible world where all items were liquid, yet the law oddly required all transactions to be priced on each side with respect to the market price of a single MOA. Say, Gold. People can trade bananas directly for haircuts, but they cannot just negotiate ratios directly, they must simply look up the gold-price of haircuts and the gold-price of bananas and offer these ratios. In that case, you still get a recession if the gold price is too high, because although bananas and haircuts can be exchanged directly, they cannot be exchanged directly at the right price. People will not trade despite being flush with media of exchange. Obviously that world isn’t going to exist at the barter limit (or maybe you don’t want to call that “barter”); I am just making a point that your paradox of hoarding post doesn’t accurately address the MOE/MOA distinction.
Antiques in that post, like rent control in other posts, are a red herring. Antiques and rents are not the medium of account. In that sense, you are right that the demand for those things or for Government Bonds, which also are neither the MOA nor MOE, is not the proximate cause of a nominal recession. But this is not in any way an effective MOE vs. MOA case. Those things are neither.
I keep meeting you on your own metaphorical terms (bananas, cows, antiques), but you won’t answer my question above. Imagine a world in which the Greek government successfully convinces people that some set of existing stock shares, say all shares that trade on the ASE, will immediately become additional media of exchange. People can now pay wages and buy coffee with shares. Every shop will have a Bloomberg terminal showing the updated Euro-value of every ASE stock. In other words, the Euro is still the MOA but now we have way more MOE. Do you agree that it is possible for people to become 100% sated in MOE, yet still have a typical recession if an increase demand for Euros purely as a store of safe savings puts downward pressure on sticky (Euro!) prices, leaving them too high?
I just realized there is an error in the grounding of the peripheral argument. We need to distinguish between science and engineering. The standard for doing something in (normal) science is: does this remove a possible falsification? The standard for engineering is: Does it work and is it useful? Medicine is not science, it’s human biological engineering.
On the Greece question, we need engineering, not science. I don’t think DSGE models are good engineering, at least not until they get specific about the time limits they are accurate for.
The analogy is to weather models, modelers know they are are not accurate past about 4-5 days, but that is still useful, and knowing that makes it more useful, not less.
“Are you accusing me of lying? ”
No, I am accusing you of exaggeration. That also goes for your claim that excess demand for the MOE is the cause of recessions. It’s like you don’t care how hard it is to be right.
More to the point, I became aware of this blog when you wrote a post about Garth Turner in which you said, “Because I was never that good at variances and covariances and CAPM and stuff.” That’s just awful – that you were never good at variances and covariances and stuff. Really, a professor of economics who can’t be bothered to understand how basic stats works? And you believe you are in a position to tell us what the cause of recessions are? You believe that even with having very shaky math skills you can entangle New Keynesian models and tells us how to reinterpret everything through the old static ISLM language?
I give this blog a hard time because the public believes the things you write and if a professor says so, it must be right. But this is such a disservice. It is so important to work through the details. It’s fine to show the public a thumbnail sketch of a complicated model, but it’s incumbent on the author to understand the model at the professional level first.
@Jim, regarding DSGE, what would you change about the NY Fed DSGE model forecast in my comment above? It shows the error bars growing with time. Doesn’t that amount to a “time limit” like you say?
Also I think there is a distinction between engineering and science (as you point out), but engineers, medical doctors, medical researchers and even car mechanics do regularly use the scientific method and thus engage in science (i.e. hypothesis formation and testing).
Also it’s difficult to engage in effective engineering (or medicine) until some basic scientific research has been accomplished. Perhaps economists recommending policy today are like a Medieval physician to the king: they are turned to in a time of acute need, but even if they want to do the best job possible, and they personally are on the right track (epistemologically speaking), they’re still left with limited (unsavory) options due to lack of advancement in basic science: leeches, cupping, the theory of the four humors and some knowledge of “medicinal” herbs. Or maybe that’s a terrible analogy. Looking in from the outside, it’s hard for me to tell.
@Avon Economic systems have non linear feedback loops, AKA they are chaotic (or complex, I understand the terminology has changed). Such systems do not have linear error behaviour. Conceivably high order polynomial, more likely exponential or worse. The uncertainties in the Fed graphs seem to me to be much too linear and small.
I could see a counter argument that says historically economies are sort of linear except for occasional large discontinuities. AKA tail risk, in that case you want to mark the timeline at a point where you calculate that the cumulative tail risk has reached a significant probability.
Yes science provides tools, but we have to evaluate them in engineering terms, whether or not a particular tool is taught in grad school is completely irrelevant.
An example from the mathematical (as opposed to the natural) sciences. Information theory has generated lots of good tools and theorems, but nobody does research in it any more. It’s done.
@dlr, not following your gold example. Why wouldn’t bananas be exchanged for haircuts if they were both too expensive in terms of gold, but can be directly traded for each other?
Nick I wonder what you make of this post by Cullen Roche.
http://www.pragcap.com/hyperinflationary-lessons-for-greece
He claims that leaving the euro was the preferred solution for Greece in 2011 but now doing it has a real risk of hyperinflation. Do you think bringing back the drachma would run this risk?
not following your gold example. Why wouldn’t bananas be exchanged for haircuts if they were both too expensive in terms of gold, but can be directly traded for each other?
nivdeita,
Because in the (absurd and thought-experiment-only) world of “barter” with an MOA, people are not really trading haircuts for bananas. The “buy” side of the transaction is acquiring an MOE — that just happens to be in the form of haircut-vouchers or bananas or whatever the buyer chooses, because almost everything is an MOE in this world. Let’s say the buyer wants a haircut and happens to use bananas as their MOE and the seller just wants to sell a haircut. They aren’t negotiating the quantities in this story, all they know is that they are acquiring or spending $10 worth of something. But the whole point of the rising value of the MOA is that the buyer no longer wants to pay $10 for that haircut, they want to pay less. They are “forced” to use the $10 MOA price as their accounting coordination system by assumption — they have no idea how many bananas/shares of Intel/golf balls they should want to trade for a haircut, so they still think they want to keep the “$10” and don’t buy the haircut* If you say this is not really barter, I agree. The point is not that this is a reasonable world or a good definition of barter. The point is just to illustrate that when Nick thinks he is isolating the MOE by assuming barter, he is actually eliminating both the MOE and MOA. So showing that a recession wouldn’t happen under barter doesn’t show that hoarding the MOE is any more proximate a cause of recessions than saving in the MOA (which of course are the same thing in most economies today).
*A side point is that to the extent nominal rigidity exists, there is surely a continuum of stickiness among stuff that would make the Calvo fairy blush. So even if you wanted to try to assume that haircut buyers would somehow make allowances for knowing that they might be paying less than $10 by paying in bananas, there would be still be large frictions as they would have no way of knowing how sticky/flexible every asset in their portfolio was relative to the stickiness of haircuts, and the sticky reference prices would still result in non-trading. The is the accounting/computing information that is crucial to the unit of account function of what we currently call “money.” The minute people can decide to trade the “right” number of bananas for the “right” number of haircuts irrespective of the reference price of $, they have eliminated both the MOE and MOA.
dlr: “Imagine a world in which the Greek government successfully convinces people that some set of existing stock shares, say all shares that trade on the ASE, will immediately become additional media of exchange. People can now pay wages and buy coffee with shares. Every shop will have a Bloomberg terminal showing the updated Euro-value of every ASE stock. In other words, the Euro is still the MOA but now we have way more MOE.”
If you take an asset (shares in your example) that is currently just willingly held (neither in excess demand nor excess supply) and magically (but that’s OK, because this is a thought-experiment) convert it into an MOE, you increase the supply of MOE but you also increase the demand for MOE by exactly the same amount. So you do not reduce the excess demand for the MOE.
But that’s not what the Greek government is doing in my example. It’s not turning an existing good into an MOE. That would only work if that good were already in excess supply (like converting Greek labour into an MOE). It is increasing the supply (from zero) of a good that is (by assumption) an MOE.
“Imagine an impossible world where all items were liquid, yet the law oddly required all transactions to be priced on each side with respect to the market price of a single MOA. Say, Gold. People can trade bananas directly for haircuts, but they cannot just negotiate ratios directly, they must simply look up the gold-price of haircuts and the gold-price of bananas and offer these ratios.”
That was exactly the world I was imagining.
“In that case, you still get a recession if the gold price is too high, because although bananas and haircuts can be exchanged directly, they cannot be exchanged directly at the right price.”
Start with all prices at market-clearing equilibrium. Now double all prices in terms of gold (the MOA). If gold is the MOE, we get a fall in trade of bananas and haircuts, with an excess supply of both goods. It’s a recession. But if we allow barter, then direct trade in bananas and haircuts can continue exactly as before, at the same relative price as before, so there is no recession, only an excess demand for gold.
My tiny macro model for microeconomists makes this clear, I thought.
Jim: in this case it’s much worse than that. Take a simple linear version of (say) the NK DSGE model. Any competent math/physics grad can solve the equations. And if they are foolishly arrogant (and some are, but not all are) they think that means they understand the model. But they don’t have a clue what’s going on. If you ask them whether this is a model of a monetary exchange or barter economy, they won’t know where to begin. Or they think that it’s totally different from the ISLM model, because the equations look different (in fact it’s just a limiting case of ISLM, with a perfectly elastic money supply function, and a coefficient of one on expected future income). Or you show them the Fisher equation, which shows a positive equilibrium relationship between inflation and nominal interest rates. And they think it means that if the central bank raises nominal interest rates it causes inflation to increase.
Nick,
Don’t both your arguments 1 and 2 depend on the assumption that there is a fixed stock of Euros? Which is certainly not true within Greece, and not even in the Eurozone as a whole.
Like if a small country on the gold standard started issuing paper money – there is a bit of seignorage revenue but other than that it I think it does basically nothing? Or in your tiny model, if the bananas can be produced at constant cost instead of there being a fixed endowment, I don’t think adding dates would do anything either.
All math depends on the premises. All models depend on some connection to the real world. I was assuming the modellers had that sorted out.
I thought the point of DSGE models was to be microfounded, any linear model is certainly not that.
It seems obvious to me that the central bank raising rates would increase inflation as a first order effect, it just gets swamped by higher order effects. If you show people linear equations and say this is how the economy works that’s impossible to see.
And (off topic, so ignore, or pick up elsewhere) how do economists deal with corrupt institutions like the current stock markets(see Michael Lewis, “Flash Boys”, or Arnuk and Saluzzi “Broken Markets”?
john: “Nick, if the greek debt is denominated in euros then wouldn’t the greek state have to pay interest on that debt (and repay the debt eventually) in euros?”
Nick: Yes.
John: “Doesn’t that imply that the greek state would REALLY prefer taxes beeing payed in euros instead of new drachmas?”
Nick: No. As long as they have the same value, at the current exchange rate, the state won’t care.
I don’t think that’s right. The Greek state has to repay its debts in Euros, not in Drachmas worth as much as the Euro debt. Euroland would not accept them. So, even if they managed to command an increasing real amount of Drachmas in tax payments, even if they managed to produce a veritably domestic boom, they would still have to borrow the Euros with which to pay back their Euro debts. Which of course is self-defeating.
The only way that a new MOE can work is if it leads to a substantial increase in exports that are paid for in Euros. They have to earn the Euros in Euroland / from Eurolanders, not at home. I don’t see that happening one way or another because there just isn’t enough stuff that Greece can export for them to dig themselves out of the hole they’re in. There’s only so many olives Euroland will eat and only so many Greek holidays they are willing to buy, especially since the rest of southern Europe is in recession too.
Or maybe I misunderstood your comment.
If you take an asset (shares in your example) that is currently just willingly held (neither in excess demand nor excess supply) and magically (but that’s OK, because this is a thought-experiment) convert it into an MOE, you increase the supply of MOE but you also increase the demand for MOE by exactly the same amount. So you do not reduce the excess demand for the MOE.
Yet this seems to be at the core of conventional monetary policy. Confused, I is.
The thought experiment I considered is that the Greek government announces a Drachma-Euro exchange rate, and then converts bank accounts to Drachma at this exchange rate.
The problem would be a self-fulfilling “run” on the new currency. Then we would expect runaway inflation (because the Euro is the unit of account), and the erosion of the value of savings. So part of the problem would be a sort of reverse Pigou effect, in which inflation erodes the value of savings (i.e. the store of value definition of money).
The obvious solution would be an exchange rate peg, but at a level that implies a substantial real devaluation. But for this to be credible, the Greek CB would need access to a lot of Euros, together with a discount lending rate to banks.
This is a very (politically) risky strategy for whoever is supplying the Euros to Greece, but if you pull it off credibly, and got the original exchange rate / interest rate right, then Greece wouldn’t actually need to use the Euros.
Nick says to dlr: “…but you also increase the demand for MOE by exactly the same amount.”
I don’t follow. Why does the demand increase?
JP: consider an economy where each individual produces one variety of fruit, and wants to consume all varieties of fruit. And there is a stock of gold that is used as a MOA and MOE. And a stock of silver that is used as jewelry. And suppose that the price of fruit is stuck too high (relative to the stock of gold), so there is an excess supply of fruit, too little trade in fruit, and a recession. But the price of silver is perfectly flexible, so the silver market clears.
Now suppose that someone invents a way to test the purity of silver easily, so that silver can be used as a MOE too. So the stock (supply) of MOE increases from gold to gold+silver. But if the existing stock of silver were already willingly held (because of the jewelry demand) at the existing price of silver, we have to add that jewelry demand to the demand for gold+silver, so the quantity of MOE=gold+silver demanded will exceed the original demand for MOE=gold, by the same amount that the supply has increased. Nobody has any extra incentive to get rid of gold+silver.
On the other hand, people would now want to switch out of gold into silver, and so the price of silver will rise, so you might argue that the real stock of MOE = (Qg/Pf + Qs.Ps/Pf) (where Pf is the gold price of fruit, and Ps is the gold price of silver) will increase, which creates an excess supply of MOE, which increases the volume of trade in fruit.
I think that’s right (though it’s not very clear).
If you take an asset (shares in your example) that is currently just willingly held (neither in excess demand nor excess supply) and magically (but that’s OK, because this is a thought-experiment) convert it into an MOE, you increase the supply of MOE but you also increase the demand for MOE by exactly the same amount. So you do not reduce the excess demand for the MOE.
The demand for the MOE means the demand to hold someone for use in exchange. Are you really arguing that if stocks suddenly became an additional MOE, that wouldn’t affect the supply/demand (i.e. relieve excess demand for) existing MOE, i.e. the demand to hold a certain amount of cash/deposits purely to use in exchange? That’s like arguing that cameras in phones won’t affect the supply/demand of existing cameras because the demand for cameras will automatically increase to offset the new supply.
Start with all prices at market-clearing equilibrium. Now double all prices in terms of gold (the MOA). If gold is the MOE, we get a fall in trade of bananas and haircuts, with an excess supply of both goods. It’s a recession. But if we allow barter, then direct trade in bananas and haircuts can continue exactly as before, at the same relative price as before, so there is no recession, only an excess demand for gold.
My tiny macro model for microeconomists makes this clear, I thought.
I don’t think this is right. Please see my reply to nivedita, above. If you are forcing people to conduct trade that at the same relative price as before, you are assuming a strange, stale MOA. But this (1) wouldn’t work, because relative prices move and if they are sticky are surely not sticky to the same degree and (2) doesn’t reflect what an MOA really means. If I am a buyer of haircuts and I know the MOA has gained value, I suddenly don’t want to pay $10 for a haircut. That’s all I know. I don’t know how many bananas to trade for a haircut and even if I do know the barber has no clue. Pure barter would allow this but again that eliminates the MOA as well. The idea that I would simply assume that even though $10 I paid for the haircut is the “wrong price” any of my MOEs would probably be the “right price” really misses the point of what it means to have unit of account in the first place. Should I know not to pay with stock, oil, wheat or milk because those prices will me more flexible relative to haircuts and so I will making a mistake?
Now suppose that someone invents a way to test the purity of silver easily, so that silver can be used as a MOE too. So the stock (supply) of MOE increases from gold to gold+silver. But if the existing stock of silver were already willingly held (because of the jewelry demand) at the existing price of silver, we have to add that jewelry demand to the demand for gold+silver, so the quantity of MOE=gold+silver demanded will exceed the original demand for MOE=gold, by the same amount that the supply has increased. Nobody has any extra incentive to get rid of gold+silver.
On the other hand, people would now want to switch out of gold into silver, and so the price of silver will rise, so you might argue that the real stock of MOE = (Qg/Pf + Qs.Ps/Pf) (where Pf is the gold price of fruit, and Ps is the gold price of silver) will increase, which creates an excess supply of MOE, which increases the volume of trade in fruit.
I don’t want to be (more) repetitive to what I just said above, but I don’t think the first paragraph is right at all. It’s not demand for the MOE that has increased. and it’s also not demand for Gold+Silver that has increased. As you said in the second paragraph, demand for silver will increase and demand for gold will decrease due solely to the increased supply of the MOE. Supply of the MOE has increased, supply of gold+silver solely as jewelry is flat, and demand for the MOE and for jewelry are presumably unchanged.
“Now suppose that someone invents a way to test the purity of silver easily, so that silver can be used as a MOE too. So the stock (supply) of MOE increases from gold to gold+silver. But if the existing stock of silver were already willingly held (because of the jewelry demand) at the existing price of silver, we have to add that jewelry demand to the demand for gold+silver, so the quantity of MOE=gold+silver demanded will exceed the original demand for MOE=gold, by the same amount that the supply has increased. Nobody has any extra incentive to get rid of gold+silver.”
This doesn’t seem right to me. When silver suddenly becomes liquid, it should now be regarded as a joint product, providing both a flow of jewelry related consumption and a flow of liquidity. (This is like dlr’s camera in phones point, where a phone becomes a joint product of photo-taking services and communication services)
This means that the stock of MOE-ness (or liquidity) has not increased from “gold to gold + silver.” Rather, the stock has increased from a sum of gold’s liquidity services to a sum of gold’s liquidity services and silver’s liquidity services.
Your conclusion that “you also increase the demand for MOE by exactly the same amount” only comes about because you’ve added silver’s quantity-of-jewelry-services-demanded to the total quantity of liquidity services demanded. But we shouldn’t be mixing apples and oranges. The total supply of liquidity services has increased while its demand stays the same.
Nick, I hope I’m not too late here, and I haven’t read the entire comment thread. My initial take is that adding the drachma would reduce the demand for the euro MOA, because the euro is also a MOE.
I agree recessions aren’t about the wrong real wage, but they are about the wrong W/NGDP. So the drachma must lower that ratio by raising NGDP (assuming W is sticky in the short run.) And you can’t raise NGDP without lowering the demand for the MOA (assuming its supply is fixed.)
Scott: Never too late! We basically agree.