Immaculate transfers and the monetary transmission mechanism

Assume that corn is identical across all countries, and is freely traded with zero transportation costs. We then know that "the law of one price" will hold for corn. The price of corn in Canada will be the same as the price of corn in the rest of the world.

Why? Suppose it weren't. First suppose the price of corn were higher in Canada than in the rest of the world. Then Canada would import all its corn, and there would be an excess supply of corn in Canada, and the price of corn in Canada would fall. Or suppose the price of corn were lower in Canada than in the rest of the world. Then Canada would export all its corn, and there would be an excess demand for corn in Canada, and the price of corn in Canada would rise.

That second paragraph is perhaps obvious, but it is not redundant. It explains why the law of one price holds.

Now suppose we have a very simple macroeconomic model of a small open economy on the gold standard, with one tradeable good - corn. To keep it very simple, assume gold is used as money. Our very simple model has only two equilibrium conditions: the price of corn in Canada equals the world price of corn; and the demand for money in Canada equals the supply of money in Canada.

Now suppose Canada discovers an extra ton of gold. What happens? Our model tells us that the new equilibrium looks exactly the same as the old. That whole ton of gold flows abroad. (Please do not argue with the model, because that is not the point of this post.)

But what is the mechanism that causes that extra ton of gold to flow abroad?

There are two stories we could tell:

1. David Hume's price-specie flow mechanism. The extra ton of gold causes the Canadian price of corn to rise above the world price, which causes Canada to import corn in exchange for gold, so the gold flows out of Canada.

2. The real balance mechanism. The extra ton of gold means there is an excess supply of gold in Canada, which causes Canadians to demand more corn, and since Canadian producers don't want to supply more corn, the extra corn is imported in exchange for gold, so the gold flows out of Canada.

The point of this post is that we cannot say that one of these two stories is right and the other is wrong. Neither is right, and neither is wrong. We need both stories; they are complements not substitutes. Someone who asserted that the first story is wrong because the law of one price always holds would be asserting that the second paragraph in this post is wrong, and that we do not need to explain why the law of one price holds. If you deleted that first story, how would you answer someone who said the second story is wrong, because the excess supply of gold and excess demand for corn in Canada would simply cause the Canadian price of corn to rise?

This is not an empirical question. The truth of the model is an empirical question, and the truth of each equilibrium condition is an empirical question. But someone who explains why the law of one price holds by explaining what would happen hypothetically if it did not hold is not asserting that sometimes it does not hold. If I explain why water in the lake has the same level everywhere by supposing it were higher in the North than South and saying it would then flow downhill from North to South you do not falsify my explanation by showing there is no empirical evidence it ever was higher in the North than in the South. I am explaining why there is no empirical evidence it ever was higher in the North than in the South.

The two stories of the transmission mechanism reflect the two equilibrium conditions. The first story assumes that the second equilibrium condition always holds, and explains why the first equilibrium condition holds. The second story assumes that the first equilibrium condition always holds, and explains why the second equilibrium condition holds.

That simple model had two equilibrium conditions and two stories of the transmission mechanism. In a model with n equilibrium conditions there would be n stories of the transmission mechanism. Or more than n, if you combined them. How many equilibrium conditions are there in the real world?

There are not two competing theories of the monetary transmission mechanism in that very simple model. There is no such thing as the monetary transmission mechanism. A "transmission mechanism" is not a causal chain; it is an explanation of why one equilibrium condition holds. The search for the transmission mechanism is a snipe hunt beloved by the people of the concrete steppes. It's a form of category mistake, like searching for which building on campus is the university.

Comments

[Mike]

"There is no such thing as the monetary transmission mechanism. A "transmission mechanism" is not a causal chain; it is an explanation of why one equilibrium condition holds. The search for the transmission mechanism is a snipe hunt beloved by the people of the concrete steppes. It's a form of category mistake, like searching for which building on campus is the university."

The transmission mechanism is just a way of describing the effect that changing the money equilibrium will have everywhere else in the economy.

[Kevin Donoghue]

If corn-production is always at the full-employment level, as it seems to be in your model, there wouldn't be much reason to discuss the monetary transmission mechanism. In fact there wouldn't be much reason to introduce money at all. You could replace gold with jewellery in this model without altering the results. A monetary transmission mechanism is more interesting if corn output can be depressed for a while, e.g. by some imperfection in price-adjustment.

[JW Mason]

The search for the transmission mechanism is a snipe hunt

.... in that very simple model. Or no?

[Nick Rowe]

Mike: "The transmission mechanism is just a way of describing the effect that changing the money equilibrium will have everywhere else in the economy."

Not really. The model's equilibrium tells you that if A changes, how B, C, and D will change. If you add 2" of rainfall to the North half of the lake, the model tells you the level of the lake rises by 1" in both halves of the lake. Now think of the millions of different ways that water could be transmitted from the North to the South half of the lake. Which of those millions of different ways is *the* right way?

Kevin: "If corn-production is always at the full-employment level, as it seems to be in your model, there wouldn't be much reason to discuss the monetary transmission mechanism."

Maybe, but economists do argue over whether Hume was right or wrong on the price-specie flow mechanism. I think I agree with the rest of your comment.

JW: not sure I correctly understand you there. Maybe I wasn't as clear as I want to be. Perhaps if I said: "the search for *the* transmission mechanism is a snipe hunt"? There are two birds (in that simple model): which one is the snipe?

[Mike]

Nick Rowe

"Which of those millions of different ways is *the* right way?"

It depends on the transmission mechanism. Who are the counterparties? How do these counterparties rebalance into assets? What is the MPC of the counterparties? How are asset holdings distributed in the economy? How accurate is everyone's info in the economy for expectations? Which entities directly hold base? etc...

[Ralph Musgrave]

Speaking as someone from the concrete steppes (whatever they are), I’m baffled by Nick’s claim that “A "transmission mechanism" is not a causal chain; it is an explanation of why one equilibrium condition holds.”

I speak English, and my definition of the phrases “transmission mechanism” and “causal chain” are identical, near enough. If Nick wants to claim they are different, can he give us definitions?

Also, speaking as a concrete stepper, I say that if rain falls in the North end of the lake, it’s gravity that equalises the level of water at each end of the lake. I call that “gravity” effect the “transmission mechanism”.

And I also claim that if that transmission mechanism weren’t there, and if there were no other transmission mechanism, the equilibrium (i.e. equal water levels in North and South of the lake) would not transpire.

Now where have I gone wrong?

[Jeff Y.]

Umm, @Mike, he was talking about actual water.

[Jeff Y.]

@Ralph, gravity is insufficient by itself as "the" mechanism of water leveling. Because gravity alone does not explain the leveling. How do the individual new rain water molecules move? Down, pushing lower molecules sideways, pushing other molecules up? Or mostly sideways sliding over the surface of the existing water? Or both? Or something else? I don't a glib "gravity" grapples with the point at all.

[Nick Rowe]

plus.google gets my point. The answer could be: one of the above; all of the above; or, it doesn't matter which of the above. I think the last answer is best, unless you are really really knowledgeable about fluid mechanics, and have actually put die in the new water and you want to follow it. But since money is fungible, there is no die in the water (bank robbers aside).

[JW Mason]

The question is whether real economies are like the lake. You are right, if we think the economy is best thought of as a market in equilibrium, then the search for a causal transmission mechanism is a snipe hunt. But if we think that economies, altho they include markets, should not be thought of as simply being markets; or if we think that markets are often out of equilibrium, so that we need to explicitly describe adjustment processes; then it is perfectly reasonable to look for a transmission mechanism. "It doesn't matter how the water got from the north side of the lake to the south side" is fine in your example. Not so fine if you then want to ask how it got to your kitchen sink.

[Ralph Musgrave]

Jeff Y,

Water molecules in the AGGREGATE cannot move sideways because of the edge or shore of the lake. Therefor rain on the North of the lake increases pressure in the water at depth X, Y and Z in the Northern half of the lake. But pressure in the Southern half at depth X, Y and Z is less. Therefor some water moves from the North half to the South half till pressures at any given depth are equalised.

I call all that stuff a “transmission mechanism” which brings about an equilibrium, namely a lake with a flat or non-sloping surface (at least on a windless day).

And please note I was brought up on the edge of Lake Windermere in the UK, and can testify to the above transmission mechanism..:-)

[Jeff Y.]

@Ralph, thanks for the additional mechanisms, but a couple of clarifications -- The "sideways" I referred to was excess north surface water sliding to the south (on the surface), And that is a simpler explanation than invoking internal pressures etc. And that's kind of the point -- jumping to one such mechanism over the other doesn't make much sense, as the actuality is that mixtures of both, and additional others, are what likely occurs. See Nick' reply about "dye".

On the other hand, I take the point, also made by Krugman, that there *is* mechanism and that its *existence* if not its details is important.

[Lorenzo from Oz]

I find papers on the Great Depression which look for transmission mechanisms without referring to what was happening with gold odd. If you do not refer to the basic framing, how can the flows of effects make any sense? Which I take to be a case in point for this post.

[Philippe]

a central banking swapping bonds for cash (monetary policy/QE) isn't anything like a ton a gold magically falling out of the sky. Is this fact relevant to your argument?

[Jeff Y.]

@Nick, sure money is fungible, but (macro) money transactions are largely, we hope, not. We don't sit down with a Mr. Gates at the poker table and say, Wow where'd he get all that moola? :-) IRS requires reporting transactions $10000 and over. Swiss bank accounts are coming under scrutiny. Of course there *is* the underground economy (pipes under the lake) to worry about...

[Philippe]

a central bank swapping bonds for cash (i.e. monetary policy/QE) isn't anything like a ton a gold magically falling out of the sky. Is this fact relevant to your argument?

[Ralph Musgrave]

Jeff Y,

Strikes me that if economics is going to be anything resembling a science, it needs to work out the exact mechanisms involved if it postulates an equilibrium.

Imagine astronomers saying something like “The Moon appears to be in a stable orbit round the Earth, so it’s obviously in some sort of equilibrium, but we aren’t too bothered what causes that equilibrium. It might be gravity balancing out centrifugal force, but we can’t be bothered checking that out.”

[Nick Rowe]

JW: It doesn't matter for my point whether or not the economy is in *market-clearing* equilibrium. Any non-market clearing model nearly always has two or more equations that determine those non-market clearing equilibrium quantities.

Philippe: it may or may not be the same. Depends on the model. But it doesn't affect my argument. The model was purely for illustration.

Lorenzo: Yep.

Ralph: you are obviously misrepresenting what both I and JeffY are saying. That is like saying we don't need a model. Jeeez!

[Jeff Y.]

@Ralph, to expand a bit on what Nick said, you would have to declare Bernoulli's work (fluid dynamics), Boyle's Law, Maxwell's equations, etc. etc. as all "nothing resembling science" because they don't explain in terms of motions/interactions of fluid molecules, gas molecules, electrons, etc.

It is a type of reiki cation fallacy: http://en.wikipedia.org/wiki/Map–territory_relation#Relationship

[Jeff Y.]

See also: http://en.wikipedia.org/wiki/Reductionism#In_science

[primedprimate]

While I agree with this post, I feel that it is important to note that focusing solely on equilibrium conditions without paying attention to the equilibriating process is dangerous. There may in fact be multiple processes occurring simultaneously and multiple frameworks through which we can describe each of these simultaneous processes, but I think we do need to have a general description of what's going on.

I am reminded of the post by Nick in which he was perplexed by Kocherlakota's claim (based on equilibrium conditions) that keeping fed funds rate low would lead to deflation. Sometimes, as Nick wisely pointed out in his post, it is essential to work through the equilibriating process. To quote Bill Woolsey's comment on that post, "I would describe the problem as too much focus on equilibrium, and not enough (Nick suggested replacing the 'not enough' with zero)on processes that would bring about equilibrium."

But perhaps I too am missing the distinction between the transmission mechanism and causal chain.

[NickRowe]

primed: I think we are very much on the same page. My problem is trying to find the right words to say what I mean clearly.

In the context of the little model I have sketched above, we can imagine two economists arguing about what is *THE* transmission mechanism, and each looking for empirical evidence to support his view. We can imagine the first economists (price-specie flow guy) trying to find evidence of cases where the Canadian price of corn was above the world price of corn. And the second economist (real balance guy) trying to find evidence of cases where there was an excess supply of gold.

And I am saying that both economists would be confused if they did this. Because: first, it's not a question of either/or, we need *both* stories; and second, because these are *counterfactual* conditional stories of what *would* happen *if* the equilibrium conditions did not hold. The whole point of counterfactuals is that we *don't* observe them.

[NickRowe]

How could we test the counterfactual conditional stories?

1. We could ban the import of corn, and watch to see if people try to get around the ban. If they don't, that would falsify the first story.

2. We could ban the sale of the new gold, and watch to see if people try to get around the ban. If they don't, that would falsify the second story.

(Maybe crappy tests, but it shows they are falsifiable in an experiment, but not from direct observation.)

[Yuriy]

I generally agree with the idea, it seems possible to express every transmissionary mechanism with a condition instead of a path. Still, if one tries to write a strict mathematical equivalent of the transmissionary mechanism condition and path, they will be the same row of equations.

Though the example seems to illustrate the same transmissionary mechanism described from the point of both gold-corn and corn-gold relative prices.
It's much more interesting when the equilibrium conditions are not the same one. For example, money emission does affects exchange rate by changing the current account as well as the capital account balance so the transmission paths are different.

[JP Koning]

Your post makes a lot of sense to me. I always thought the debate between the concrete steppes people and yourself could be boiled down to a focus on theory vs institutions. The concrete steppes people are concerned with the tangible aspects of monetary policy like open market desks, legal departments, and discount windows, whereas market monetarists are more interested in top level theory. The concrete steppes folks want to know how theory translates into the nuts and bolts of a real live central bank, sort of like how the physics behind internal combustion might be translated into an actual car engine by engineers. But perhaps there is more to the debate than I've been able to pick out.

[Akshay]

Nick,

As insightful as ever. I have a few doubts though - maybe from a "concrete-steppes" perspective:

Assume:
(1) More corn = more happiness for the people (with VNM utilities and all that). Gold is a way to buy corn and store "purchasing power" to buy corn in the future - but gold does not bring any additional happiness besides that of providing a convenient way to buy corn.

(2) People have a maximum limit on how much corn they can consume per day. That limit has been gradually increasing throughout the centuries due to advances in corn-consumption technology.

(3) People can however use a small store-room in their homes to store a finite volume of corn or gold.

(4) Corn is susceptible to pests/enzymes/sunspots which cause it to rot after a random amount of time. The time of decay is a random variable with some variance/standard deviation/volatility.

So, say somehow the volatility of corn-decay increases - new variety of pests/unseen seasonal variations etc. People, not knowing when their existing store of corn will rot away, substitute it with gold. Gold becomes super-pricey - people are not lending it any more to each other. Some super-duper authority entrusted with managing "gold" in the economy injects more of it into the system to let people buy time to figure out a solution to the pestilence.

So, my questions:
(a) Would increasing the supply of gold solve the problem in a convincing way? Or would better methods to estimate time-of-decay and making the corns more resistant to pests/decay be more logical solutions to bring corn consumption to the original level?

(b) If the super-duper authority does not inject gold itself in the system but it tries to do so via intermediaries who like to lend to people who seem to have "better" corn (as collateral) - does increasing the gold supply do any good if there *is* no "better"/less-volatile corn out there?

Which leads to my point - does the transmission mechanism really not matter?

[doudoune moncler femme]

What's up, after reading this amazing post Worthwhile Canadian Initiative: Immaculate transfers and the monetary transmission mechanism i am also delighted to share my knowledge here with colleagues.