How many monetary transmission mechanisms are there?

An economic historian builds a macroeconomic model to help her understand the gold standard. Her model says the central bank sets the dollar price of gold, and the stock of base money is demand-determined. If the central bank raises the dollar price of gold, her model says this will cause an increase in the general price level, and an increase in the stock of money.

She thinks about the monetary transmission mechanism in her model, how an increase in the dollar price of gold will affect the demand and supply of gold, the demands and supplies of close substitutes to gold, radiating out to the demands and supplies of all other goods.

Then the economic historian decides to study the debate between advocates of the gold standard and advocates of the silver standard. She builds a new model to help her understand this debate. Her new model contains both the price of gold and the price of silver. Under the gold standard, the central bank sets the price of gold, and the price of silver is an endogenous variable. Under the silver standard, the central bank sets the price of silver, and the price of gold is an endogenous variable.

Under the gold standard, if the central bank raises the price of gold, that will cause a rise in the price of silver. Under the silver standard, if the central bank raises the price of silver, that will cause a rise in the price of gold.

How many monetary transmission mechanisms does her model contain?

It is tempting to answer "two". There is one monetary transmission mechanism under the gold standard, which explains how changes in the price of gold affect the price of silver and other endogenous variables. And a second monetary transmission mechanism under the silver standard, which explains how changes in the price of silver affect the price of gold and other endogenous variables.

But then consider:

If the central bank wanted to, and if it responded quickly enough to shocks, the central bank could adjust the price of gold to hit any desired target for the price of silver.

If the central bank wanted to, and if it responded quickly enough to shocks, the central bank could adjust the price of silver to hit any desired target for the price of gold.

How could an outside observer even tell the difference? Is the central bank adjusting the price of gold to get the price of silver where it wants it to be? Or is it adjusting the price of silver to get the price of gold where it wants it to be? Which price is the instrument, and which price is the target? Would it make any difference to any individual agent in the economy which of those two things the central bank is doing?

If the two monetary policy transmission mechanisms are observationally equivalent, and if the two monetary policy transmission mechanisms make no difference to any individual agent in the economy, does it make any sense to say there are two different monetary policy transmission mechanisms in her model?

If, despite this, you say that her model contains two different monetary transmission mechanisms, one from the price of gold, and a second from the price of silver, and that the central bank can choose to use either one, then consider:

There are millions of different goods in a real world economy. There is the price of gold, the price of silver, and the millions of other prices of all the millions of other goods too. If the economic historian's model contains two different monetary transmission mechanisms, then the real world must contain millions of different monetary transmission mechanisms.

So why are so many monetary economists fixated on just one of the millions of possible monetary transmission mechanisms? One which starts from the nominal rate of interest on very short term loans between banks? Aren't they being just a little narrow-minded by looking only at one, when there are millions?

You want concrete steps? I will give you a million different flights of concrete steps.

For example, like Lars Christensen, lets talk about the monetary transmission mechanism where the central bank adjusts the stock price index, rather than a short term nominal interest rate. No zero lower bound problem there.

(This post is a reflection on Josh Hendrickson's recent post.)

Comments

[Ritwik]

Because it (the short rate) is the only price that enters (in mainstream thought) ALL inter-temporal decisions. Of consumption, as well as wealth maximization. The other prices are all relative prices.

As you must know, the prices that enter even the very short term loans between banks are not all the same. Even on the same day, different banks borrow and lend at different rates. So even this *price* is actually multiple prices. It's not that particular market price that matters the most - it is the point that that price is used (supposedly) as a reference for all other decisions.

How effective this mechanism is then depends on how widespread it is used as reference, or how much the institutions and processes around this reference(term structure of expectations, levels and distribution of wealth, efficiency of financial markets etc.) are stable. So yes, it is often ineffective when left for the cold as a reference(as Greenspan found when long rates refused to budge even as he hiked the short rate repeatedly) or counter-productive when used for the wrong reason (as in fuelling commodity/ housing/ carry trade booms in an attempt to manage liquidity and preserve intermediation). But the macroeconomic stability tool of a monetary authority has to necessarily answer the question of - which instrument coordinates inter-temporal decisions, in the aggregate, most optimally/widely. And the short rate is usually the right answer.

And yes, stock price indices can be targeted. That would maintain stability, for sure. We should just be prepared for a world where every firm becomes Enron/GE and tries to be as correlated (and informationally insensitive) to the market index as it can, with the only flights of entrepreneurial fancy in imagining various new ways of using cash to make some more cash, with investment bankers being the preferred suppliers as well as consumers.

[Min]

"If the central bank wanted to, and if it responded quickly enough to shocks, the central bank could adjust the price of silver to hit any desired target for the price of gold.

"How could an outside observer even tell the difference?"

Well, lessee, how would the central bank adjust the price of silver? It could buy or sell silver, or silver futures, or silver stocks, right? It could announce the price at which it would buy or sell silver.

You would think that an outside observer would notice. ;)

[Nick Rowe]

Min: "It could announce the price at which it would buy or sell silver."

"Announce"? You mean its "communications strategy"?

Just like the Bank of Canada announces its new target for the overnight rate, and the actual overnight rate just jumps to the new target, without the BoC actually doing anything (so the BoC guys tell me)?

I wonder how much gold central banks actually bought and sold under the gold standard? They certainly didn't keep 100% gold reserves. how much gold did the Fed buy when FDR raised the dollar price of gold in 1932?

Ritwik: "Because it (the short rate) is the only price that enters (in mainstream thought) ALL inter-temporal decisions. Of consumption, as well as wealth maximization. The other prices are all relative prices."

But that is just repeating the puzzle. Why does mainstream thought think that way? Because that's how it thinks of the monetary transmission mechanism.

[Nick Rowe]

Min: the Bank of Canada also has gold reserves. Sometimes it even buys and sells gold reserves, at various prices. But nobody pays any attention. Nobody sees this as any sort of signal of monetary policy. All people wonder about is whether it's making a profit or loss. Whether it might be prudent to keep more gold for some unspecified emergency. Maybe gold miners pay attention. Macroeconomists don't.

Hmmm. There might be a post topic here.

[Ritwik]

Nick

I believe it's not repeating the puzzle. You are thinking of the fundamental macro-economic challenge as one of optimizing/ co-ordinating same-period exchange. I submit that the fundamental problem is of coordinating/ optimizing decisions over multi-period tradeoffs.

Inter-temporal/ exchange prices may still be out of whack even if the Wicksellian price is right. But if the Wicksellian price is wrong, they have no hope of being right. You can argue whether directly targeting the Wicksellian price is a good way of getting it right, or if a better way exists. But you have to recognize that the aim is still to get that price right, leaving the rest to a decentralized system.

Monetary transmission mechanisms dominate all else, so that whatever you choose will enter all markets.
But they have to be subservient to the price of time, which exists and behaves independently.

[Nick Rowe]

Ritwik: Dunno. I think it cuts both ways. If real interest rates are wrong, it's going to be hard to get intra-temporal allocation right. But if intra-temporal prices are wrong, it's also going to be hard to get inter-termporal investment/saving decisions right. Lipsey second best sort of arguments.

Also, even if I did buy your argument that the interest rate is *the* most important relative price. Does it necessarily follow that monetary policy should use that price? If monetary policy gets that price wrong (which it will) maybe that would have far worse conequences than if monetary policy got some less important price wrong?

I'm thinking back to my old post.

[Min]

Moi: "It could announce the price at which it would buy or sell silver."

Nick Rowe: "Announce"? You mean its "communications strategy"?

"Just like the Bank of Canada announces its new target for the overnight rate, and the actual overnight rate just jumps to the new target, without the BoC actually doing anything (so the BoC guys tell me)?"

Well, I am skeptical of the Chuck Norris theory, but I will play along for the sake of argument. ;)

Nick Rowe: "I wonder how much gold central banks actually bought and sold under the gold standard? They certainly didn't keep 100% gold reserves. how much gold did the Fed buy when FDR raised the dollar price of gold in 1932?"

I followed this blog for some time, until he stopped posting: http://newsfrom1930.blogspot.com/

Ships used to carry gold back and forth between countries quite frequently as their CBs bought and sold gold. :)

[JoeMac]

Nick,

Is the Keynesian interest rate mechanism the same thing as the Monetarist excess cash balances mechanism? I've never been able to see how these two are one and the same. Where is the hot potato in the story of the interest rate mechanism?

[Nick Rowe]

JoeMac: they certainly sound very different, don't they? Just as the transmission mechanism from gold price to silver price will sound different from the transmission mechanism from silver price to gold price; after all, they start and end at opposite places? But can they really be different? If you had a model in which gold and silver prices were linked, would you ever be able to tell if gold caused silver or if silver caused gold? Or are they co-determined simultaneously, so that our linear stories are just linear stories about an (Artsie) non-linear world??

[JoeMac]

Well, I've always thought of them as different. The Cash balance effect seems to move from reserves to an increase in M and then directly to NGDP. But the interest rate mechanism seems to move from an increase in reserves to interest rates (or tobin's q) and then to Velocity, and then to NGDP. In the cash balance mechanism the same dollar that is added by the CB is spent on goods. But in the interest rate mechanism (and tobin's q) the dollar that increase interest rates (and equity prices) is not the same dollar that is used to purchase goods. I think that's because of the distinction between primary and secondary markets in financial asset markets that doesn't exist in goods markets.

[rsj]

Nick, I am confused by what you are trying to show here.

Are you really making an argument that the interest rate mechanism is identical to the quantity-of-reserves mechanism -- in which case, I would like to see the argument.

Or are you arguing by analogy, saying that because you can concoct two mechanisms that are equivalent, that all other mechanisms must also be equivalent?

[Nick Rowe]

rsj: Let me spell it out a bit more explicitly.

Suppose you agree that, in a model which contains the gold price and the silver price, it doesn't matter whether the central bank sets the price of gold or the price of silver; there is only one transmission mechanism in that model, not two.

OK, now let me add the price of copper to that model. The central bank can also set the price of copper. If you agreed there is only one, not two, you should presumably now agree that there is one, not three.

Now add a fourth good to the model. Now you should presumably agree that there is one, not four.

And so on, up to a million goods, including bonds, and commercial bank IOUs. Now you should presumably agree that there is still one, not one million.

Either you agree there is one, or one million. Or, I think, the onus is on you to say that something is different somewhere along the argument.

[rsj]

Nick,

I would say that a commercial IOU price is not the same as the price of gold or copper.

If I buy copper from you, then you do not have the copper any more, I have it. But when I take out a loan from a bank, the bank gives me its IOU and takes my IOU. There is no limit on the number of IOUs that can be so created.

Similarly, when I repay a loan, I give the bank its IOU (deposit) and it extinguishes my IOU (loan) held by the bank.

Show me how you can costlessly create gold or copper, or how you can costlessly destroy gold or copper in a way similar to the taking out of a loan or repayment of a loan, and I would be more amenable to this argument.

Here, you may find SRW's blog to be interesting:

http://www.interfluidity.com/v2/3402.html

[Nick Rowe]

rsj: there is no limit to the amount of paper the central bank can print to buy gold or IOUs. There is a limit to the amount of gold I can and will produce to sell to the central bank. There is a limit to the number of (trustworthy) IOUs I can and will produce to sell to the central bank.

[rsj]

The central bank, if it incurs a capital loss, must push that loss onto the rest of government, so it is borne by taxpayers.

You have a similar limit to credit expansion on the part of the CB and private sector balance sheets. Both entities are concerned with capital losses.

And, btw, it is for this reason that democracies prevent the CB from purchasing whatever asset the CB wants, just as bank regulators prevent banks from purchasing whatever IOUs the banks want.

[rsj]

Or maybe the point of your post is to argue that CBs, as currently constructed by both law and custom, are too conservative in expanding their balance sheets, and we should change the controls on central banks just as we change the controls on private sector banks.

I think that would be an interesting politco-economic discussion.

Generally speaking, the purchase of risky assets is construed as fiscal policy and so central banks are loathe to purchase IOUs not guaranteed by the government. Once the government steps in and provides loan guarantees to the private sector borrower, the fiscal policy is assumed to be done, and the CB can then purchase the guaranteed asset without being accused of conducting fiscal policy by itself. That avoids the issue of absorbing defaults.

In any event, the "monetary space" is finite for central banks and currently consists of the space of government guaranteed securities, which is much smaller than the space of all securities.

[Nick Rowe]

rsj: consider these four monetary policies:

1, Buying and selling gold to hold the price of gold fixed.

2. Buying and selling silver to hold the price of silver fixed.

3. Buying and selling gold to hold the price of silver fixed.

4. Buying and selling silver to hold the price of gold fixed.

If you were forced to group those 4 policies into two pairs so that members of each pair were most similar in their macroeconomic consequences, how would you pair them?

I would pair {1 and 4}, and {2 and 3}.

"Or maybe the point of your post is to argue that CBs, as currently constructed by both law and custom, are too conservative in expanding their balance sheets, and we should change the controls on central banks just as we change the controls on private sector banks."

That wasn't the point of my post.

And, oh God, it's about *shrinking* the balance sheets of some central banks (like the Fed). You are thinking like a hydraulic Keynesian. Expected deflation, and fear of continuing recession, is what makes the Fed's balance sheet so abnormally large.

[rsj]

Nick,

I don't think that central banks can or ever did control the price of gold by buying and selling gold.

They controlled the price of gold by setting interest rates.

An investor would form some belief about the future price of gold, take into account the storage costs, and the central bank would set interest rates in such a way so that the joint solution of this equation is such that the price of gold is constant. Or it goes up/down.

Gold, silver, are long lived assets. They are not consumption goods.

The prices of long lived assets are completely determined by the discount rate, rental yield and expected re-sale price. It doesn't matter who else owns the asset.

I.e. if Nick owns half the shares of IBM (congratulations!), then that does not change my valuation of IBM, or how much I am willing to pay for IBM stock, or how much anyone else is willing to pay for it. If Nick sells his holdings of IBM, then still I don't care. The price of IBM does not rise or fall based on how much of it is owned by Nick, and how much is owned by other people. Neither does the price of gold rise or fall based on how much is owned by the central bank.

The price of gold vis-a-vis dollars changes based on the expected return for holding dollars (i.e. the interest rate) versus the expected return for holding gold (storage costs, any rental yield + expected resale price).

[rsj]

And the reason why central banks during the gold standard era held stocks of gold was not because they needed to hold stocks of gold in order to set the price of gold, but because they needed to hold stocks of gold in order to perform their reserve banking roles.

Gold was used to settle international trade flows -- it was a type of reserve global currency -- and this meant that nations needed a small positive amount of gold in order to be able to absorb variations in trade flows. The only time the quantity of reserves became important was when nations saw net gold outflows due to trade, i.e. due to their currency being overvalued. In that case, the CB needed to raise interest rates or de-value, and until that happened, it needed enough gold in its reserves to cover the outflows without destroying its domestic banking system.

But the quantity of gold held by a central bank was never a function of the price of gold in terms of that nation's currency.

[Min]

Nick Rowe: "1, Buying and selling gold to hold the price of gold fixed.

"2. Buying and selling silver to hold the price of silver fixed.

"3. Buying and selling gold to hold the price of silver fixed.

"4. Buying and selling silver to hold the price of gold fixed.

"If you were forced to group those 4 policies into two pairs so that members of each pair were most similar in their macroeconomic consequences, how would you pair them?"

FWIW, I would group 1 and 2 together and 3 and 4 together, because 3 and 4 introduce more uncertainty.

BTW, didn't the US go off of a dual metallic standard because the attempt to keep the price ratio of gold to silver constant did not work? (At the time it had no central bank, though.) Any attempt to control the price of one metal by controlling the price of the other would be similarly fraught with difficulty in the real world, no?

Nick, does this post relate to your previous one about simultaneity? If you assume simultaneity the practical problems disappear. ;)

[Old Shep]

Nick,

According to the "Report on the Management of Canada's Official International Reserves"
at the Department of Finance "There were no gold transactions during the period, with the last of the Government's gold bullion holdings having been sold in December 2003"

It seems the extent of Canada's gold "reserves" is comprised of gold coins and other bullion products offered by the Royal Canadian Mint at the retail level.

Apparently in 1980 the BoC made the decision to divest itself of its gold holdings.By December 2003, they were gone. Zero. Nobody seems to know why Canada bailed on gold, or least nobody is talking about it. A number of entertaining conspiracy theories make for intriguing speculation.

[Nick Rowe]

Min: "BTW, didn't the US go off of a dual metallic standard because the attempt to keep the price ratio of gold to silver constant did not work?"

I don't know, but that sounds very plausible to me. There have been various attempts in history, IIRC, to fix the relative prices of gold and silver in a bimetallic system (Isaac Newton was put in charge of the Royal Mint and fixed the ratio at x??). Bimetallic standards fail. You can see why. A central bank can only fix one variable, and it must be a nominal variable. By trying to fix two nominal variables (the dollar price of gold *and* the dollar price of silver) it is also fixing a real variable (the relative price of silver to gold). Bimetallism failed for exactly the same reason that the attempt to target "full employment" failed in the 1960's and 1970's. Both were attempts to use monetary policy to target a real variable. (It can succeed for a time, but the central bank will have to continuously buy silver and sell gold (or vice versa) until it runs out of gold and has massive reserves of silver. It's a bit like trying to keep the price of wheat permanently artificially high by buying wheat. The government has an ever-growing wheat mountain.

"Nick, does this post relate to your previous one about simultaneity?"

Good question. It does relate, but I can't (yet) clearly explain how it relates.

Old Shep: Thanks. Damn. Now you say it, I remember that I once knew that, but had forgotten it.

[Min]

Hmmm.

If the CB, say, buys gold to bring its price up to $X and lets the price of silver adjust, won't the price of silver be lower than if the CB buys silver to bring the price of gold up to $X? In the first case the demand for gold increases more than the demand for silver, and in the second, the demand for silver increases more than the demand for gold. The ratio of the prices of gold and silver is not fixed. If the CB is buying gold, it is not buying silver, and vice versa.

When the US went off of the dual standard in 1873, the price of silver in US$ dropped by half.

[W. Peden]

The US went off a bimetallic standard because of their Civil War- the fairly standard practice of abandoning a commodity standard during a major war, so as to partly finance the war through inflation. The Fourth Coinage Act in 1873, which was the effective adoption of a gold standard by the US (the legal position was confusing until 1900) was AFTER the US had had a fiat currency for years.

The consequences of adopting the gold standard instead of a bimetallic standard were so disastrous that the US government was forced in 1878 into a humiliating u-turn, with the Bland Allison Act in 1878 legally obliging the US Treasury to buy silver at 16:1, i.e. at very close to the old ratio of 15:1. This pseudo-bimetallism existed from 1878-1893, but it's worth noting that this was NOT a bimetallic standard- it was a peg, rather than a commitment to exchange gold and silver on demand at a fixed ratio. So, when the US Treasury's commitment was removed after the Panic of 1893, it was abandoning a gold-silver peg rather than a bimetallic standard; the peg failed because of a fall in gold reserves to the legal limit.

The debate in 1893 was very similar to the debate today in some respects: one side argued that a deflationary monetary policy based on a gold standard with no silver peg would restore confidence and allow a recovery, while the other side argued that an inflationary monetary policy based on (some form of) retaining 16:1 would enable a recovery.

Like any price ceiling, a bimetallic standard creates distortions i.e. Gresham's Law meant that silver disappeared from circulation as its nominal value exceeded its real value. However, Friedman argued in "Money Mischief" that the interchangeability of gold and silver under a bimetallic standard was stabilising. I can't remember his argument, though.

[J.V. Dubois]

"The price of IBM does not rise or fall based on how much of it is owned by Nick, and how much is owned by other people" No, but the price of IBM does depend on how much money and for how long Nick and other people are willing to hold in mattresses, affecting velocity and therefore prices. And central bank is the biggest player in this area so all people need to watch closely how CB behaves if want to have their price expectations right.

Anyways Nick has a very good point here. It is really strange that economists are so fixed on the short term interest rate zero lower bound. In this article: http://www.themoneyillusion.com/ Scott comments on idea that central bank could change their target from short term interest rate (that is at 0) to long term interest rate. It is still an interest rate, a word that is familiar to modern macroeconomists, but it is far from zero bound. The transmission mechanism is the same: we are willing to expand the balance sheet until we see this interest rates fall down.

[Nick Rowe]

Min: "If the CB, say, buys gold to bring its price up to $X and lets the price of silver adjust, won't the price of silver be lower than if the CB buys silver to bring the price of gold up to $X? In the first case the demand for gold increases more than the demand for silver, and in the second, the demand for silver increases more than the demand for gold. The ratio of the prices of gold and silver is not fixed. If the CB is buying gold, it is not buying silver, and vice versa."

Forget about what the CB is buying. Anybody can buy gold and silver; that's not what makes the CB special. Think about what the CB is buying it with: freshly printed money. That's what makes the CB special. That's where the CB gets its power to influence the prices of gold, silver, and everything else.

Yes, if the CB buys a large percentage of the total stock of gold, or silver, or wheat, or whatever, and puts it in the basement and out of public hands, that has a real effect. Just like the wheat mountain. But that's not where the action is.

The CB can make the price of wheat converge to either zero or infinity. Not by making the supply of wheat either infinite or zero. But by making the supply of money either infinite or zero.

[W. Peden]

The funny thing is that the "only consider short-term interest rates" approach has become so common that I've found it quite hard to explain Keynes's liquidity trap to some economic students, since for Keynes it's long-term rates that are most important. I'm not sure how Hawtrey triumphed over Keynes in this debate (he certainly hasn't got the credit) but even Keynesians don't seem to think like Keynes when it comes to interest rates.

(There is the additional problem, in explaining Keynes's liquidity trap, that economic students today in the UK aren't accustomed to thinking of the quantity of broad money as an exogenous variable. I don't think there is any School of Keynesianism that wouldn't give Keynes an "F", at least without doing some very dubious reinterpretations of his exam answers.)

So, quite apart from making life difficult for macroeconomists, the short-term interest rate thinking makes life harder for studying and teaching the history of economics.

[Nick Rowe]

W Peden: "This pseudo-bimetallism existed from 1878-1893, but it's worth noting that this was NOT a bimetallic standard- it was a peg, rather than a commitment to exchange gold and silver on demand at a fixed ratio."

Can you explain that a bit more please? In particular, what does "peg" mean here?

JV: Yep. I think there are only 3 possible answers to the question "How many monetary transmission mechanisms are there?": Zero, it just goes bang bang from one equilibrium to the other; One, the monetary hot potato is the only fundamental thing that central banks can do, and anything else is just a gloss on the hot potato story; millions.

I have noticed, in arguing with some non-economists, and MMT people, etc., that they can only ever focus on what it is that banks and central banks are buying, not on what they are buying it with. So they can't see the difference between monetary and fiscal policy. It all looks fiscal to them.

[W. Peden]

Nick Rowe,

A peg here means to commit to open-market operations in order to influence the nominal price such that an exchange rate is fixed or within a certain range. So the standard example is a currency peg e.g. the ERM system, where this is carried out on the basis of purchases. This can be compared with strict convertability, where the government is committed to exchanging any quantity of A for B at a fixed price.

In the 1878-1893, the Treasury would buy silver within a range to keep the market price at 16:1. There was no legal obligation for the US government to give you gold or silver at 16:1 when requested.

[Ryan V]

rsj: "I.e. if Nick owns half the shares of IBM (congratulations!), then that does not change my valuation of IBM, or how much I am willing to pay for IBM stock, or how much anyone else is willing to pay for it. If Nick sells his holdings of IBM, then still I don't care. The price of IBM does not rise or fall based on how much of it is owned by Nick, and how much is owned by other people. Neither does the price of gold rise or fall based on how much is owned by the central bank."

What if Nick has A) the authority to issue new IBM shares and B) enough cash to credibly commit to buying IBM shares at any price? That seems like a better analogy for me.

Central banks only ever indirectly set the price of gold. It's easier to think of them setting the price of money in terms of gold. And they could do that because they have A) the printing press/asymmetric redeemability (which lets them lower the price of money in terms of gold) and B) sufficient gold reserves to buy money at any price (which lets them raise the price of money in terms of gold).

They key insight in Nick's post, I think, is that the central bank could also buy money with silver to set the price of money in terms of gold. How is that monetary policy different from buying money with gold?

[Nick Rowe]

W Peden: Got it. Thanks. I should have figured that's what you meant.

Right now Canada has a (sort of) CPI peg, but the BoC does not make the CPI bundle of goods convertible on demand into Loonies.

[Nick Rowe]

W. Peden: Yep. In other words, we are talking about what is sometimes called "indirect convertibility".

Ryan: "What if Nick has A) the authority to issue new IBM shares and B) [a printing press hidden in his basement] to credibly commit to buying IBM shares at any price? That seems like a better analogy for me."

There, fixed it for you! ;-)

"They key insight in Nick's post, I think, is that the central bank could also buy money with silver to set the price of money in terms of gold. How is that monetary policy different from buying money with gold?"

Yep. It's not different (except for possibly very slight changes in the composition of the asset side of the BoC's balance sheet, which are peanuts compared with world stocks of gold and silver). So how can the transmission mechanism be different? Even though one transmission mechanism starts with silver and the other starts with gold? They both start with money. That's what matters.

[W. Peden]

Nick Rowe,

In that respect, I'd be willing to talk about the existence of a CPI peg, but not a CPI standard i.e. the dynamics are presumably more comparable to an exchange rate peg than a commodity standard like strictly defined bimetallism.

[Phil Koop]

1) There are a million different mechanisms. Because, even if each has identical effects on the total price level, each will have a different effect on the relative price of some commodity which has a real economic use, and therefore each mechanism has a different real effect.

2) Notwithstanding 1), interest rates really are special. Because, bonds are universally required by every economic agent and have no substitute.

3) But technically, there aren't really a million prices. Because, the total nominal value and price elasticity of the chosen commodity instrument must be such that there is no credible possibility of either the CB buying all of the commodity in existence or selling all it has before its target is achieved. For example, if the CB chose original works of art by a minor artist as its instrument, it might well find itself unable to control the price of gold (or whatever.) This is true even though with a better-chosen instrument, the CB would not actually have to buy or sell (very much of) it. That is the true point of the People of the Concrete Steps: you can't influence expectations without a credible policy instrument.

[rsj]

What if Nick has A) the authority to issue new IBM shares and B) enough cash to credibly commit to buying IBM shares at any price? That seems like a better analogy for me.

Then Nick is free to buy all of IBM. But it's not going to change how much I value IBM.

[rsj]

On the other hand, the central bank can change the rate of interest by buying and selling a tiny amount of reserves, effectively zero reserves, and that *would* change my valuation of IBM.

One is an effective channel, whereas the other is not. To change interest rates, the CB need only add a bit of reserves, leave them there for a day, and then pull them back out, and the rate of interest is permanently at the new value desired by the CB. To try to move the price of equities, it has to rely on illiquidity in the asset markets, purchase trillions of dollars of equities (which is illegal, btw) and then after a while the illiquidity effects go away and the price of equities returns to what is before.

It's pretty clear that there is a real difference between these two channels, at least to me. One works and the other doesn't.

[Nick Rowe]

rsj: "Then Nick is free to buy all of IBM. But it's not going to change how much I value IBM."

If Nick Has a printing press it will. Because you value IBM in terms of dollars, and by printing more or fewer dollars Nick will change the value of the dollar. If I double the price of IBM shares in terms of dollars, I halve the value of dollars in terms of IBM shares.

Actually, it's the interest rate channel that doesn't work. We always knew that as a theoretical possibility, because theory says that if I double the number of dollars and halve the value of each dollar, then all variables with $ in the units will double, but the rate of interest does not have $ in the units, it has units 1/years. And now, at the ZLB, we have learned that that channel doesn't work in practice either.

[Nick Rowe]

Phil: 1. The only effects of different mechanisms on relative prices is if they affect the composition of the asset side of the Bank of Canada's balance sheet. If it controls the price of silver the BoC may hold some stocks of silver; if it controls the price of gold it may hold some stocks of gold. But those stocks are miniscule relative to total world stocks of silver and gold.

2. "Because, bonds are universally required by every economic agent and have no substitute."

How does the old saying go? "Neither a borrower or a lender be"? Family, or church, can be a substitute. So can equities. What about Islamic finance?

How about the price of food as a monetary standard?

3. "Because, the total nominal value and price elasticity of the chosen commodity instrument must be such that there is no credible possibility of either the CB buying all of the commodity in existence or selling all it has before its target is achieved."

What about with indirect convertibility (what W Peden calls a "peg")? Or even less than 100% reserves? The CB doesn't need to buy and sell gold, let alone hold 100% gold reserves, or even marginal 100% reserves, to peg the price of gold.

Sure, it's a bit trickier to peg the price of an illiquid good, because price information is so poor.

[rsj]

Nick, you are confusing the unit of account ("dollar") with the medium of exchange, deposits and currency. The cb, we're it allowed to purchase shares of IBM, would not cause the quantity of deposits or currency to increase. It would cause the quantity of reserves held by the banking system too increase. This is what we've seen happen on both the US and in Japan.

While I agree that setting short rates to zero doesn't always work, we know that QE never works, or at least there has not been an example of it working. That just means that monetary policy doesn't always work in setting the price level to what you want.

[Nick Rowe]

rsj: "While I agree that setting short rates to zero doesn't always work, we know that QE never works, or at least there has not been an example of it working."

I don't like the term "QE". It's a silly new name for something we already had a name for, and that central banks have always done. "Open Market Operations (OMO)" is the normal term. Actually, about 95% of what central banks have always and ever done is OMO.

Central banks buy and sell things. And they communicate their future plans for buying and selling things and the conditions (including prices) at which they will buy and sell things in future.

What are the things that central banks buy and sell?
1. IOUs signed by the government. That's OMO (aka QE). That's 95%(?) of the balance sheet of central banks in normal times). That's gross tuning, in Canada.
2. IOUs signed by commercial banks. Reserves and stuff. Fine tuning.
3. Forex. Mostly only relevant for fixed exchange rates. Peanuts otherwise.
4. Gold. ditto.
5. Distressed assets in emergencies. Lender of Last Resort stuff. Not part of normal monetary policy for targeting nominal variables.
6. Paper, ink, economists, computers, etc.

If "QE" never works, it is incomprehensible how the BoC has managed to keep inflation almost exactly (on average) equal to the 2% it said it was targeting over the last 20 years. Ditto for all the other inflation targeting countries. And for gold price targeting countries before that.

[rsj]

Nick,

QE refers to quantity adjustments in the stock of money rather than interest rate adjustments. To see the difference, note that when a CB conducts OMO, it conducts reversible operations that permanently change the interest rate. I.e. you add $100 million of reserves and the short rate starts to fall, say from 5% to 4% to 3%. If you left the $100 million of extra reserves, the short rate would drop all the way to zero. Instead, when it hits 3%, if 3% is your reserve price target, then you take the $100 million back out.

Therefore from start to finish the quantity of reserves changed was zero. And yet the price of reserves adjusted downward from 5% to 3%.

That's what CBs normally do. They exploit the relationship between the deviation of reserves supplied from that of reserves demanded and the rate of change of the marginal cost of reserves.

The result of exploiting this relationship is that for the same quantity of reserves, you can have any interest rate. There is no relationship between levels.

What you are describing, with QE, is an attempt to exploit a relationship between the level of the interest rate and the level of the quantity of reserves.

Such a relationship between levels does not exist, has never existed, and so cannot be exploited, nor has it ever been successfully exploited.

[rsj]

Sorry, I forgot to close the bold tag. ugh.

[Fixed - SG]

[Nick Rowe]

rsj: "What you are describing, with QE, is an attempt to exploit a relationship between the level of the interest rate and the level of the quantity of reserves."

No. What I am describing, with OMO, is an attempt to exploit a relationship between the price level and the level of the quantity of base money. Any (negative) effect on interest rates is purely temporary, and due to sticky prices, and the effect of higher base money on higher real income and higher saving. (And the (positive) relationship between nominal interest rates is with the rate of change of base money.)

[rsj]

No. What I am describing, with OMO, is an attempt to exploit a relationship between the price level and the level of the quantity of base money.

No, the OMO exploits a relationship between the quantity of bank reserves and the rate of change of the marginal cost of reserves.

That is why, after each of the last rounds of QE as done by the Fed, the quantity of "money" -- deposits and currency -- remained unchanged, while the quantity of excess reserves increased dollar for dollar with increase in the CBs balance sheet.

The quantity of deposits and currency is demand determined to first order by expenditure/income smoothing needs of households, and to second order by interest rates. Therefore once the short rate hits zero, additional balance sheet expansion by the CB has no effect on the quantity of non-financial sector money in the economy.

[rsj]

http://research.stlouisfed.org/fred2/graph/?g=92A

[Min]

Nick Rowe:"Forget about what the CB is buying. Anybody can buy gold and silver; that's not what makes the CB special."

No, but we were talking about discernable differences. I think that a 50% drop in the price of silver in 1873 is quite discernable. (OC, that was a special, one-time event. ;))

The question is not whether the CB can, in the extreme, control the money supply. The question is whether different ways of affecting the money supply are all same-same.

[J.V. Dubois]

RSJ: "Therefore from start to finish the quantity of reserves changed was zero. And yet the price of reserves adjusted downward from 5% to 3%"

This is really a quite silly stuff. Adopting your line of thinking, central banks conduct monetary policy by words. Today, Mario Draghi said the following two sentences "Within our mandate, the ECB is ready to do whatever it takes to preserve the euro. And believe me, it will be enough" and markets suddenly shot up. This has to be how the economy really operates, right?

The fact that the central bank can do the trick that you describe is entirely due to the expectations channel. I assure you that if Central Bank would not have credible target (such as inflation) then expanding balance sheet and reserves of banks would not cause nominal interest rates to fall. Or in other words, if central bank manipulates nominal interest rates without any change in the money quantity, it had to influence money velocity. Only other explanation is that the CB controls real interest rate, which is complete nonsense.

[rsj]

J.V.,

This is really a quite silly stuff

It is not silly at all. But your comment points to a huge disconnect between what central banks do, and what economists think that central banks do. The economists would be better served by studying how the institutions actually operate.

Adopting your line of thinking, central banks conduct monetary policy by words.

Exactly. They can move the overnight right merely by making an announcement, which is what happens in practice.

The fact that the central bank can do the trick that you describe is entirely due to the expectations channel

No, it is because the banking system needs a certain amount of reserves to absorb the expected volatility between inflows and outflows. If it does not have them, then it cannot, in aggregate, get more. If the banks in aggregate need $100 of reserves, but only $99 are available, then they will bid up the price of reserves and keep bidding it up. The alternative to not having enough reserves is to declare bankruptcy. If they need $100 and get $101, then there will always be $1 of excess, one bank will lend it out, but then another bank has the excess, so it lends it out. They will keep bidding down the lending rate until it reaches zero.

This is really simple stuff, but because it doesn't match some archaic quantity based supply and demand fantasy, economists refuse to believe it, regardless of how often they are told by the people who do this stuff for a living that this is how things work. For example, take a look at "Modern Central Bank Operations" by S. Fulwiller:

http://www.cfeps.org/ss2008/ss08r/fulwiller/fullwiler%20modern%20cb%20operations.pdf

[rsj]

Or in other words, if central bank manipulates nominal interest rates without any change in the money quantity, it had to influence money velocity.

Yes, that is the only operative channel available to the central bank. It must encourage households/firms to create more money and invest or spend the proceeds somehow, bidding up prices. Which is why in some cases the central bank cannot control prices, if firms/households don't want to borrow at the current risky rates available to them, given that the short risk-free rate is zero.

[J.V. Dubois]

rsj: "If they need $100 and get $101, then there will always be $1 of excess, one bank will lend it out, but then another bank has the excess, so it lends it out. They will keep bidding down the lending rate until it reaches zero"

This is the prime example of mechanistic, "concrete steps" approach to the monetary policy. You probably think that Because this is how things seemed to work last few years, it has to remain so under any circumstances. Examine your example in two extremes:

1. What will happen if CB prints it just 1 dollar of excess reserves? In your example, banks collectively get rid of this dollar, only CB can sterilize it therefore if CB is unwilling to do so, this $1 it has to cause fall of interest interest rate to 0 for the whole economy, right? Let me try to provide some explanation for this nonsense. What if banks collectively know that lowering interest rate will cause inflation to shoot up, which in turn will cause people willing to hold more cash in their wallets, which means that people will demand more paper cash as opposed to money on the reserves, which will in turn make some bank to just hold that one dollar of "excess" reserve to meet this demand.

So we end up by having excess dollar absorbed by public in terms of larger demand to hold currency to offset whatever inflation increase printing this one dollar of excess reserves has. Simple expectations driven hot potato effect

2. What will happen if CB prints trillion dollars tommorow by buyng government debt and promises never to revert this?

[Philo]

Really, the transmission mechanism is the same in both scenarios; the difference lies in the starting point of the transmission—the initial action by the monetary authority that has to be transmitted to the broader economy. In the gold-standard scenario, the monetary authority’s initial action is to buy or sell gold, not silver, and also to announce its peg for gold, not for silver. In the silver-standard scenario, it is the reverse. (Of course, when the authority is pegging the price of gold, and doesn’t care about the price of silver, the price of gold will be steady while the price of silver fluctuates randomly; when the authority is pegging the price of silver, it is the reverse.) The difference between the two scenarios—between the gold standard and the silver standard—is not a difference in *mechanism*.

[Sergei]

J.V. Dubois, it is not silly and not non-sense. One dollar might not be able to move rates that much but only because daily "mistakes" of banks in aggregate are much larger. That is why there are *always* excess reserves. But if you, the central bank, exceed the tolerance level of commercial banks for "mistakes" in their liquidity management with your reserve operations, then rates will go down. And they will go down until the tolerance of banks for mistakes will reconcile with your new reserve injection. If you inject 1 trln then that it is sure way to push rates all the way down to zero.

Finally, to change the rate all the central bank has to do is to announce it. And it is done. No OMOs are required for this. Liquidity optimization of banks does not really depend on the level of interest rates but it depends on spreads.

[wh10]

JV- would you rather incur a cost of 0% (i.e. not holding onto the excess reserves now and borrowing it when rates go to 0%) or incur an opportunity cost of something above 0% (i.e. holding onto the excess reserves instead of lending them out at some rate above 0%)?

[J.V. Dubois]

Sergei: "If you inject 1 trln then that it is sure way to push rates all the way down to zero."

This is utter nonsense. Imagine an economy with inflation of 2% and nominal interest rate of 6% - so the real interest rate is 4%. Now CB announces that it increases the inflation target to 100% they will buy whatever it takes by newly printed money to make it so. So if I am correct, you say that CB flooding commercial banks with cash causes nominal interest rates to fall to 0% because "that is how the banking sector operates". So just like that CB will lower the real interest rate from +4% to -100%. And in the meantime all owner of assets that CB decided to buy - for instance pension funds formerly holding government bonds - will be happy swapping these assets for quickly detoriating currency.

Come on guys, this is just plain stupid. Go read something about Fisher Effect.

[rsj]

What if banks collectively know that lowering interest rate will cause inflation to shoot up, which in turn will cause people willing to hold more cash in their wallets, which means that people will demand more paper cash as opposed to money on the reserves, which will in turn make some bank to just hold that one dollar of "excess" reserve to meet this demand.

Why would banks "know" something that isn't true?

What a bank knows is that is has $1 of excess reserves earning 0 interest. Regardless of what economic theory you happen to believe, $1 of reserves that you don't need, earning 0 interest, is inferior to a $1 overnight loan earning positive interest. Any positive interest above zero is preferable.

Therefore the bank will lend that dollar. Now, another bank has an extra dollar. It will also prefer to lend the dollar, etc.

This is pure profit maximization.

Now CB announces that it increases the inflation target to 100% they will buy whatever it takes by newly printed money to make it so.

There are laws that prevent the CB from purchasing "whatever it takes". Of course, if the CB agreed to purchase everyone's labor for $1000/hour, many would quit their jobs and work for the CB, and I guarantee you that inflation would shoot up. That is fiscal policy. But inflation would shoot up because you've increased everyone's income, not because you've increased the quantity of reserves.

Given the universe of things that the CB is allowed to buy -- government guaranteed assets -- the CB cannot credibly commit to targeting something that it can't control. And while the CB certainly has influence over inflation, it does not precisely control it, as can be seen from our own experience, or the experience of Japan.

[J.V. Dubois]

wh10: No, what I say is that if printing money in whatever amounts causes nominal interest rates to go down, then there would be queues of people before banks willing to lend for this 0% buying whatever real assets they can and then repay the debt it later when the inflation hits and pocket the profits.And banks also know it, suddenly there is plenty of high quality potential lenders who may make huge profits. Not only that, there will be a huge number of people attacking banks and changing their deposits for cash because they will not be satisfied by 0% interest when they can get rich by buying real assets. We could end up with bank runs. If banks don't want to lose all deposits they will have to increase their interest rate on deposits in line with expected inflation.

[rsj]

J.V.,

I am confused by your above paragraph -- did you mean to replace "lenders" with "borrowers"? How are people going to attack banks -- you mean violence against banks, the threat of which will prevent banks from lending too much or too little?

But yes, the general idea is that lower interest rates cause people to borrow more and invest more and this causes an increase in the price level.

If we call this the Wicksell channel, then what I and others believe is that this is the only operative channel available to the CB, and that the quantity channel is a null op.

Some nations, such as Canada, have a zero reserve system, so they are clearly able to control interest rates while keeping reserves (quantity) fixed at zero (actually, it is something like $50 million due to the "frictions" that Sergei talked about).

We know that, at the ZLB, increasing the CBs balance sheet merely causes reserves to go up, it does not cause inside money to change at all, nor does it cause the quantity of currency to change, or the quantity borrowed to change, or NGDP change, etc.

There are many papers out there (I referenced one) describing how these institutions operate and how rates are set, and we have the practical experience of Japan and the U.S. as well.

[Sergei]

J.V. Dubois, your depiction of banking is very strange. Bank liquidity is generally a scarce asset. Central bank makes it so but running its monetary policy. That is central bank generally regulates the cost of liquidity. That is the whole point of monetary policy. In different systems central banks use different tools. In developed countries this tool is price of liquidity, i.e. some tenor on interbank yield curve. In the USA it is O/N, in eurozone it is 2 weeks, in Switzerland it is 3 months and so on. Central bank does not set the *market* rate at this tenor on the yield curve. It rather declares its *desired* level. And if market forces move the market rate from the targeted level, then central bank has to intervene. What you effectively call non-sense is that if central bank intervenes into the market by the tune of 1 trln dollars, this intervention will have no effect on the market. Well, what you say is generally non-sense though I can imagine a situation (like Sep-Nov 2008) where even 1 trln would not make the trick.

Whether people buy assets or not, banks make profits or not, inflation goes through the roof or not, there is a line of borrowers or not, it all does not matter at all. Interbank market does not care about everything you mentioned and much more. It cares only about the rate from today to tomorrow. And when tomorrow comes, it will care about it again. Simples.

And as I said above banks do not care that the level of rates is. Ie. there is no difference to banks whether rate is at 100% or 2%. Banks care only about spreads which are generally defined by the competition. However, when banks make loans, they do care about *future* rates because it is their risk factor. Which might completely negate the argument you are trying to make regarding profitable borrowers etc.

[J.V. Dubois]

rsj: Yes, I meant "to borrow" and I may have used the word attack incorrectly. My excuse is that I am not native speaker, while in truth it is just my laziness and that I do not double check what I write every time. Sorry for that. Also, replace "whatever it takes" for "whatever amount of assets CB can legally buy". The last I recall there is $14 trillion of US public debt, there is plenty of room to go before FED will hit the wall.

Ok, so back to your example: imagine that all banks have the right amount of reserves (or that they fulfill all capital adequacy ratios in countries where reserves are not required). As it happens, there is $1 of excess CB money in the system. My first question is - why should any bank need to borrow this $1 from whatever bank happens to hold it now? What can be their expectations other than lend it to yet another bank at s loss with even lower overnight interest rate?
The only possible impact this $1 of excess reserve has is expectations. All banks know that $1 of excess reserves means that they collectively can make another $10 of loans in the future (if capital adequacy/reserve ratio enforces such money multiplier). But if all people who could get their loans with yesterday's interest rates already have one, the banks will just need to attract new customers. One way to do it is to offer lower interest rates so that the excess reserves will be absorbed. So the interbank interest rate will not get down to 0. It will fall by whatever amount that equalizes one-day opportunity cost of holding cash without gaining any interest on it with long-term profits for the bank if they will be able to issue $10 debt tomorrow for a given costs of securing it on interbank sector. (ok, they will also need to attract new deposits, but set this aside for the moment as it is not important for this example) The first bank to find such marginal customer in their databases of recent loan applicants will borrow and hold these excess reserves so that they can send the customer a new offer.

Sergei: All the central bank does is that it prints money. Therefore, CB can influence nominal price of whatever they want - if CB credibly commits to it. So your claim that commercial banks do not care about inflation expectations is equal to the claim that commercial banks do not care about monetary policy. Any bank that would care only for today's interest rate would quickly go out of business.

"However, when banks make loans, they do care about *future* rates because it is their risk factor. Which might completely negate the argument you are trying to make regarding profitable borrowers etc."
No, this actually completely supports the argument that I am so painfully trying to make. I will give you yet another example: if nominal interest rate would be 0% and FED would announce that it wants 100% inflation within the year, this would have massive effects on expectations:

1) When banks offer loans, they try to assess the future income of potential lender as well as future price of his collateral. With expectations of 100% inflation and 0% interest rate both these variables would be highly advantageous to lenders. Almost anyone would be eligible for bank loan. Banks could get massive profits even if lenders go bankrupt just by selling the collateral at new inflated price.

2) When bank offer loans they need to take into account the funding. One of the crucial features of banking is that it borrows short and lends long, so they need to be careful with their balance sheet. With 100% inflation and 0% nominal interest rates offered for deposits, banks would face imminent risk of bank runs by depositors. Banks are just facilitators: they are intermediaries between depositors and lenders. So if 1) is true and almost anyone could make profit by such conditions, that invariably means that every depositor is at risk of huge loss.

Combine 1 and 2 and you will find out that for banks to remain in business, to prevent bank runs, they need to substantially increase the interest rate they offer. Despite of CB pumping insane amounts of liquidity to the system, the demand for this liquidity would increase proportionaly. This is the key finding for you: it is not banks that set the interest rate on the market. It is the depositors and lenders that do this. Banks serve only as intermediary, truing to gauge how the future can look like, asses the risk associated with this future and make profit on this.

Both: I know that you are not going to be convinced here. But for your own sake, start thinking about all you say. I know that cognitive dissonance can be very painful to overcome, but the positions you hold are absolutely indefensible. The only way you do it is to constantly mire yourself into mechanics of how banking sector in current operational setup seems to operate trying to extrapolate this into almost insane arguments.