From gold standard to CPI standard

Inflation targeting is a much better monetary policy than the gold standard. But that's not what this post is about.

Is there any fundamental theoretical difference between how monetary policy worked under the gold standard and how monetary policy works today for a modern inflation-targeting central bank? That's what this post is about.

If you believe that monetary theory fundamentally changed, or needed to change, when modern central banks abandoned the gold standard, then this post is for you.

[Update: on thinking it over last night, and reflecting on my own beliefs over the years, I should probably add that the question in this post ought to be directed at myself, as well as directed at others. I had always thought "Oh yeah, dropping gold changed everything". But other than making the basket much more representative and sensible, which matters massively in practical terms, did it really change things theoretically?]

1. Start with a monetary system where the central bank's paper money (or an electronic equivalent) is 100% backed by gold reserves, and each note can be redeemed at a fixed price for gold.

2. Now relax the assumption of 100% gold reserves. The central bank holds fractional gold reserves, with the rest of its assets being something else, like government bonds. If there's a demand to redeem notes, the central bank might have to sell bonds to redeem some of the notes, if it fears it might run out of gold reserves. And interest rates might rise if it does this.

3. Now take the limit of that system, as the percentage of gold reserves approaches 0%. There is indirect convertibility of paper money into gold. The central bank never actually buys or sells gold, but it buys or sells bonds to keep the equilibrium price of gold pegged, so that people can still swap their notes for gold in the open market at a fixed price.

4. Now replace the fixed peg with a crawling peg. The central bank allows the paper currency to depreciate at 2% per year against gold. The target price of gold rises at 2% per year.

5. Now replace that target price of gold with a target band. The price of gold can fluctuate either 1% above or 1% below that target price, at the central bank's discretion. The exchange rate of paper money for gold is a snake in an upward-sloping tunnel.

6. Now allow some base-drift. If the price of gold rises towards the top of the tunnel, the central bank may re-base the tunnel upwards. If the price of gold falls towards the bottom of the tunnel, the central bank may re-base the tunnel down. But any such re-basing is unbiased, so the expectation is always that the price of gold will rise at 2% per year from now on.

7. Now add silver to the gold. The central bank targets a weighted average of the price of gold and the price of silver. There's a synmetallic standard.

8. Now add a lot of other goods and services to the basket, along with gold and silver.

9. Now remove any goods and services that are not included in the CPI basket of goods, and weight the remaining goods according to their weights in the CPI.

[Update: Maybe add 10: Since some goods in the CPI basket have sticky prices, that need time to change, the central bank switched to pegging the expected future price of that basket.]

We now have a modern inflation-targeting central bank.

(For a price level path targeting central bank you can delete step 6.)

At what point in the above steps did monetary theory need to fundamentally change?

This post was inspired by the second of two very good recent posts by Edward Harrison. (The first is here.)

Comments

[Jon]

In steps #2 and #3 the transmission mechanism changed.

[Nick Rowe]

Jon. At step 2 we are definitely still on the gold standard. Central banks did not have 100% gold reserves under the gold standard.

But that does raise the important question: even under 100% gold reserves, what is the transmission mechanism? If the price of gold were to start to raise, and the central bank sells gold to bring the price of gold back down, why does the price of gold fall? Is it because: there's more gold in public hands?; there's less money in public hands?; a bit of both?

[HJC]

Under 100% reserves isn't there a no-arbitrage condition. There is no need to buy gold from a non-CB seller with a higher price than the CB; and vice versa. As we move from 100% this gets weaker.

[Nick Rowe]

HJC: I don't really see the difference. In 2, if people start to redeem their notes for gold at the CB, the CB starts selling bonds to remove more notes from public hands, so it doesn't run out of reserves. 3 is just the limiting case of this process, where as soon as one person starts to walk towards the CB to sell his notes for gold, the CB sells enough bonds until that person changes his mind and buys bonds instead.

[HJC]

My point was in response to "If the price of gold were to start to raise". It is hard to understand why someone would pay more for gold than the CB was offering at. Perhaps I should have asked: price of gold rises in terms of what? Please ignore the 'weaker' comment for now.

[Nick Rowe]

HJC: OK. I meant if the price of gold started to rise in terms of the central bank's notes. But you are correct; it would only start to rise if the central bank were a little bit slow to redeem the notes. Perhaps I should have said instead: "If there started to be an excess supply of notes relative to gold, so that the price *would* start to rise unless the central bank did something to eliminate that excess supply".

[HJC]

First, do we agree that there cannot be an excess supply of notes under 100%?
Then for the sub-100% case, I think we are beginning to look at the relative pricing of bonds and gold in terms of notes, since some notes are now backed by bonds. If the note-prices of bonds and gold start to diverge, then the CB might have a problem.

[Nick Rowe]

HJC: "First, do we agree that there cannot be an excess supply of notes under 100%?"

If the central bank is open 24/7, any excess supply would be immediately eliminated. It might take till Monday morning otherwise. But it would be the same under 2 and 3. It's just that under 2 and 3, the central bank could sell bonds as well to eliminate any planned or incipient excess supply of notes.

[HJC]

I'm confused as to who is issuing the excess notes. Surely, by definition all notes (and reserves) are backed by equivalent gold amount. Isn't that what 100% means? Are you talking about private bank credit? Only when we move to sub-100% can there be any excess. What am I missing?

[nemi]

Isn´t this a strange question?

Essentially, a fiat currency simply has more degrees of freedom than one that is backed by e.g. gold, i.e. the possible actions/strategies under a gold standard is (at least almost) a subset of those possible under a fiat currency.

Off course, most Austrians/libertarians would not agree. They would say that while that might be technically true, human nature will make sure that the ones in control of the fiat currency is “misusing” it. This might be true if some set of commitments are credible under a gold standard, but are not under a fiat currency – but I do not see why that would be the case to any large extent (since you always can drop the gold standard).

But, I believe that there are very large difference in practice, since I do think most CB´s do make use of the additional flexibility of the fiat currency.

[nemi]

PS: I do think you are right here, but I do not think it is correct to demand that someone who disagrees with you should be able to point out exactly where they disagree (maybe I am wrong, I do not have any knowledge about rhetoric)

Similarly to people showing why libertarians actually want to eat babies, or that taxation is the same as slavery, you use something similar to the Sorites paradox (but even harder to argue against since you have qualitative as well as quantitative changes). Even if you can’t defend the difference between young/old, few/many, pornography/non pornography etc., that does not mean that there are no difference or that the distinctions are meaningless.

[J.V. Dubois]

HJC: I thinks Nick is right here. Imagine CPI targeting as a regime where CB backs-up its currency 100% by all goods and services produced in its monetary domain. Redeemability is enforced by government law that everyone has to accept the legal tender.

And another note - flow (and thus also stock) of gold may vary even under 100% reserves. Gold can be mined, it can be imported or exported, it can be consumed by industry. I see no fundamental difference between money being backed up by gold or money being backed by goods in CPI basket. If growth in Asia permanently increases demand for gold jewelry then CB needs to tighten to protect the price of the dollar. Under CPI regime, if growth in Asia increases demand for CPI goods than CB also needs to tighten to protect the price of the dollar (as CPI goods flow out of the country and inflation rises).

The main difference here is that in gold regime tightening (in form of money stock contraction) occurs automatically since money is destroyed at the very moment when gold exporter redeems his dollars for gold. In CPI basket regime, the tightening occurs indirectly, the money base is sterilized by CB selling government bonds to the private sector.

The point is, that it does not really matter if CB holds the thing it targets (gold or CPI goods) in warehouses so that it can manipulate the prices of these things directly. What matters that in the end the "transmission mechanism" is just the manipulation of money base. If in gold price targeting the price of gold increases too much, that means that asians will have to pay horrendous amounts (in nongold goods and services) to obtain it. It is the money base contraction that ultimately stops the outflow of gold out of the country. The same is valid for CPI goods targeting.

[HJC]

J.V. Dubois: Thanks for your reply. I was specifically wondering how there could be an excess supply of notes in the 100% gold-backed case, as Nick's reply to me implied.

[genauer]

I think it is pretty clear. The situation changes in the moment, when the CB does not longer guarantee to exchange notes against a fixed amount of gold.

This is between step 2 and 3, also you carefully avoid to spell out "guarantee to exchange notes against a fixed amount of gold". As in every good government game :- ) Never tell the hard truth, always use some pleasant words.

So, this was 1973, End of Bretton Woods.

[Bill Woolsey]

With a gold standard that keeps reserves (100% or fractional) changes in the central bank's holdings of gold influences the equilibrium price of gold. Changes in the quantity of money cause changes in the prices of other goods and services, which brings the relative price of gold into equilibrium.

If the amount of reserves held by the central bank is constant (not in percent but in absolute terms,) which would always be true with a zero reserve gold standard, then there is no possibility of manipulating the equilibrium relative price of gold by the central bank changing its own demand for gold. It is only the change in the quantity of money (presumably by purchases and sales of bonds) that bring other prices in line with the equilibirum relative price of gold.

In either case, the equilibrium relative price of gold depends on supply due to mining and demand for industrial purposes, mostly jewelery, and also demand for monetary purposes. In the no reserve system, the demand for monetary purposes is awlays zero. If the absolute amount of reserves were held constant, it would neither dampen nor cause changes in total demand.

So, the difference is the change in the demand for reserves impacts the relative price of gold.

If we use a medium of account where storage is impractical, then this never has an effect, only the effect of changes in the quantity of money changing the actual relative price of the medium of account applies.

In my view, the real divergence is between a gold standard where we imagine all the gold in the world is money, and its quantity is approximately fixed. There is a bit of mining and industrial use, but we can ignore that. Paper money augments this quantity of "base" money. Well, it isn't base, but rather precious :). We can come up with various multipliers between the amount of monetary gold and the quantity of money available.

And, of course, we can split this up nation by nation. How much of the world gold is in each nation, and then there is some multiplier to determine how much money is in each nation has. And that determines the price level in each nation. If the national price levels are out of balance, gold flows between nations, generates changes in the quantity of money in each nation, and leads to an adjustment in each nation's price level.

And then, we compare that to a world where gold is a commodity mined out of the ground and used to produce jewelry and dental work, and so on. Oh, and some of it is used for monetary purposes as wel, but that gold is made into coins or stored in vaults is an after thought. Money is mostly debt instruments, priced like everything else, in terms of gold. The monetary ones typically have fixed prices, which requires the issuers to limit their quantities to the amount demanded, otherwise they would trade at a discount or premium from their face values.

If there are different "countries" all using gold as the medium of account, then the price level in all the countries depends on the world supply and demnad for gold. If there are different central banks, then each of them limits its issue of paper money according to the demand to hold it. Generally, gold goes from mines to the industrial users. Central banks can manipulate the world price level a bit by changing the amount of gold they choose to hold. Offseting an increase in the demand for gold requires ample gold reserves. Accumulating gold reserves is easier, but a "small" central bank would have little impact on the world gold market.

In this second world, price level targeting is like replacing gold with a bundle of commodities. The economics isn't that different. And then, switching to inflation targeting or nominal GDP targeting isn't that different either.

On the other hand, the first version of the gold standard seems quite different than targeting the CPI. And 100% reserves really have nothing to do with it. It is whether we can treat the world stock of gold as money, with mining and industrial uses as an afterthougt, or we can treat mining and industrial uses as determining the relative price of gold, with monetary uses as an afterthought--and the price of gold defining the unit of account. Gold coins and gold reserves make a difference between it impacts these two magnitudes.

Sometimes I want to say that both framings of the problem are equally good. But in truth, I think the "all gold is money," national gold holdings, money multipliers, and the like, is just the wrong way to look at things.

With the gold as medium of account framing, we can think about what happens when gold is coined. When gold is used as reserves. When we put on proportional gold reserve requirements. We can think about switching the medium of account. We can do devaluations and revaluations.

And to return to another debate, that paper money is a type of debt falls out naturally from this framing. Treating paper money as paper gold comes from the alternative framing. Instead of a national quantity of gold reserves, we have paper gold. Rather than being imported or mined, it is "mined" at the mint. Just like gold miners spend gold, the government spends the paper money. With the other framing, paper money is a type of debt. If issued by a government institution it is government debt. Swtiching the medium of account to a bundle of goods doesn't change that.

[Chmee]

I'm failing to understand one point here. If currency was 100% backed by gold (or silver, or some other commodity that doesn't have dramatic supply swings, or even a combination thereof), why would the price of those commdoities ever rise at all? Unless the CB decides to change the ratio, which they shouldn't be allowed to do, the price is fixed, isn't it?

I would further argue that we wouldn't, and shouldn't even need a CB if we were on a fixed commodity standard. The whole premise of a CB with people running them that think they are smart enough to manage the economy hasn't exactly worked out too well, has it.

[Ryan]

I think the change pretty obviously occurs at 2. Here's what Mises had to say (citation: http://mises.org/humanaction/chap17sec11.asp and http://mises.org/humanaction/chap17sec12.asp)

If the debtor--the government or a bank--keeps against the whole amount of money-substitutes a 100% reserve of money proper, we call the money-substitute a money-certificate. The individual money-certificate is--not necessarily in a legal sense, but always in the catallactic sense--a representative of a corresponding amount of money dept in the reserve. The issuing of money-certificates does not increase the quantity of things suitable to satisfy the demand for money for cash holding. Changes in the quantity of money-certificates therefore do not alter the supply of money and the money relation. They do not play any role in the determination of the purchasing power of money.

If the money reserve kept by the debtor against the money-substitutes issued is less than the total amount of such substitutes, we call that amount of substitutes which exceeds the reserve fiduciary media. As a rule it is not possible to ascertain whether a concrete specimen of money-substitutes is a money-certificate or a fiduciary medium. A part of the total amount of money-substitutes issued is usually covered by a money reserve held. Thus a part of the total amount of money-substitutes issued is money certificates, the rest fiduciary media. But this fact can only be recognized by those familiar with the bank's balance sheets. The individual banknote, deposit, or token coin does not indicate its catallactic character.

The issue of money-certificates does not increase the funds which the bank can employ in the conduct to its lending business. A bank which does not issue fiduciary media can only grant commodity credit, i.e., it can only lend its own funds and the amount of money which its customers have entrusted to it. The issue of fiduciary media enlarges the bank's funds available for lending beyond these limits. [p. 434] It can now not only grant commodity credit, but also circulation credit, i.e., credit granted out of the issue of fiduciary media.

I would add that I think point #5 seems a little suspicious to me. The price of gold is not very volatile, especially in comparison with the price of a fiat currency. Therefore, the price of gold isn't really fluctuating in your example, it's an illusion. It is only the exchange ratio that's fluctuating. And if one side of that ratio is relatively stable (i.e. gold) and the other is comparatively unstable (i.e. fiat currency), then it is obviously only the fiat currency that is fluctuating.

It's not impossible that the price of gold fluctuates, but in your example, the fiat currency is ever-rising, and the price of gold is not. Perhaps I've missed the point of 5.

[K]

Nick,

Things become a bit unclear at step 2. When you go from 100% to fractional reserves, do you also modify the capital structure of the CB? The 100% reserve money can be thought of as equity. But when you move to fractional reserve, do you also introduce an equity liability of the CB? If not, at what step do you make this critical change in the capital structure of the CB, which largely delinks the asset value from the money value.

I think it stops making sense step 8. You can't "back" money with
consumption goods. You can only use capital goods (including possibly things like houses). Stocks and bonds are fine, but not haircuts. Consumption goods are a flow, not a stock. The impossibility of "storing" consumption is why money inevitably fails as a consumption guarantee. The only way to defer consumption is to invest in capital goods. That's why money should be backed by capital goods.

[wh10]

Nick, I posted this at Ed's. I think the transmission mechanism fundamentally changes when you change from a convertible regime to free floating. I think changes from quantity of reserves to interest rate on reserves.

"Nick, I am admittedly not at all sure about this, but what are your thoughts on this?

under a convertible system, the ratio of qty of reserves to the thing being exchanged has a direct impact on the exchange rate, right?

but is that the same thing as the ratio of qty of reserves to the price level under a nonconvertible system? i struggle to see why because, even if i go along with your policy endogenous interest rate conceptualization, it’s still a story about the price or interest rate on reserves directly affecting appetite for credit, as opposed to qty of reserves.

I think that’s roughly what Scott was saying here:

“I would argue that it might look that way but the causation is wrong. The CB is targeting inflation via an interest rate target rule, which affects bank lending (customer desires to borrow, obviously), and then the demand for reserve balances at the target rate at any point in time, which the CB accommodates.”

Disagree? (Probably :).)

[JP Koning]

Very good post.

"At what point in the above steps did monetary theory need to fundamentally change?"

I don't think it ever did. But I guess that depends on what your definition of monetary theory is.

It seems to me that you could substitute the "central bank money" in your story with any other private liability (or even equity) and the story is still true. A private company could issue bearer notes initially 100% backed and convertible into gold and then go through all the permutations you've listed to arrive at a note linked to CPI.

Does your definition of monetary theory include the lender of last resort? Because the original 100% bank couldn't be a lender of last resort until it made the switch to fractional.

[Bill Woolsey]

Under a competitive gold coin standard, seingorage is based upon the cost of producing coins. Mines earn the bullion price of gold and industrial users pay the bullion price of gold. It is free to adjust based upon supply and demand--within limits. If there is a hip hop fad and an increase in the demand for bling, the first effect is a decrease in the demands for whatever it is the bling buyers would have purchased instead, along with an increase in the demand for bullion to produce the bling. The price of bullion rises, and the demand for some variety of goods and services fall. Those who would have sold those goods have less money, and reduce their expenditures. It is likely that some firms respond to being short on money by selling short term securities, which lowers the prices of those securities and results in higher interest rates. To the degree there are deposit banks, taking short term deposits and making loans, the shortage of money proper could result in a withdrawl of those funds and so fewer bank loans and higher interest rates on those loans.

At the same time the usual deflationary process is occuring, less newly mined gold shows up at the mint to be produced into coins. The coins needed for trade (which is falling already) are withdrawn and notes are retired. If the price of bullion rises to the point where it is equal to the gold content of coins, jewelers can redeem notes for bullion, assuming the note issuers hold it. So, the quantity of gold money falls, freeing up gold for bling.

The higher interest rates, lower real output, and lower prices of assorted goods and services all reduce the demand for money, and the demand for gold for industrial pruposes, including bling. While perfect neutraility is unlikely, for the most part, the long run result will be that the prices of goods and services, including resources like labor, fall enough so that the production of all goods, including gold, return to productive capacity, and interest rates fall again to the natural interest rate where saving equals investment. At the lower price level (and wage level) mining gold is more profitable at past levels, and so there is more mining. More gold goes into bling. The real quantity of money is higher meeting the demand.

How does this compare to zero reserves? Rather than the liquidity effect on interest rates only occuring as those short on money sell securities or pull money out of "savings" accounts, it happens right off as the increase in the bullion price results in the "central bank" (or private banks) selling off liquid securities to keep the price of gold rising to the redemption point. (Which would have to do with what it costs to mint gold coins, though they no longer really exist.) If the price of bullion gets to the point where jewelers go to the bank (s) rather than usual dealers, then the banks, having no gold reserves, have to buy it on the market and sell at a lower price. But, those buying the bling still spend less on other things as before. That, and the higher interest rates generated by the bank(s) result in less spending in total, lower prices and wages. In the end, it is more profitable to mine gold. The amount of paper money is lower (the banks sold off bonds and contracted the quantity.) The demand for gold is lower because its relative price is higher. This partly offsets the initial increase in bling demand, reduces other industrial demand. Overall production and interest rates are back to normal.

If the government runs the system and has "free coinage," then there is less (or no) room for variation in the market price of bullion. At first pass, I would expect the jewelers to start melting down coins to make bling. (Wasteful, no?)

If it is a bullion stanard at the price is literally pegged, then I guess jewelers get free transport. Just pick up the raw materials from the bank. (This is a reason why I doubt a free banking system would allow unlimited redemption at a fixed price. I know I wouldn't want to receive lower interest payments on my money so that jewelery is cheaper to produce.)

In reading comments here by fans of the gold standard, I notice that they seem fixated on the one thought experiment of an excess supply of paper money at an unchanged relative price of gold. While that could happen, it hardly exhausts the possibilities.

[Scott Sumner]

Nick, Great post, I did something similar back in 2009. :)

[Edit. I embedded Scott's link here NR]

[Bill Woolsey]

Imagine two scenarios:

In the first, the banks hold no gold reserves, but still promise to pay gold on demand. There is "slack" in the system because they used to promise to pay gold coins, but they don't exist now. Long ago, the banks promised bullion but at a price slightly higher than the defined dollar price of gold, reflecting the cost of the old disused coins.

In the second, the banks promise to provide securities with a current market value equal to the price at which bullion would have been provided. The banks generally keep these interest bearing securities, and it is better from their perspective than gold. Also, banks don't want to provide storage and transport for the bullion market.

It seems to me there is no real economic difference between direct convertibility and indirect convertibility.

Now, suppose that there is no legal convertibility, but if the price of bullion rises above a small allowed premium (or, if you prfer, the notes trade at more than a small allowed discount,) the bankers face criminal pentalties. They are fined or maybe face jail time.

As in scenario one (where they promise to provide bullion,) and scenario 2 (they promise to provide securites with a market value equal to the bullion,) their is a rule constraining the bankers to adjust the quantity of money they issue so that is value is fixed relative to gold.

Maybe one kind of incentive structure is more effective than another, but that is what all of it is about. Rules to keep the issue of paper money limited to the demand to hold it at a price in terms of the medium of account (gold.)

[dwb]

At 6, we are off the gold standard i would say, in terms of thinking about the reaction to an increase in the demand for gold (due to liquidity preference).

At step #2, if there is an increased demand for redemption (liquidity) then the CB sells bonds (how do we know there will be enough bonds around?) to buy gold, which raises rates. This potentially chokes off investment, reducing AD, which further increases the desire for liquidity to the extent people believe loans will go bad.

Seems to me that since the CB has only a finite ability (number of bonds, which are backed by what-taxes?) to defend the price (debt::gold) then the price/peg/band in 3,4,5 is no longer 100% credible. It could fail. Especially if, as Bill Woolsey notes, there are other countries and a world market.

At 6, the bank is esentially agreeing no longer to fix the $/gold redeemability.

So suppose there is now an increase in the demand for money (liqidity preference, as above). in #6 the bank has the freedom to devalue, and hence it does not need to sell *any* bonds. interest rates are not impacted, and so neither is AD. in #6 we've gained a degree of freedom.

[J.V. Dubois]

Bill: Excellent post. "In my view, the real divergence is between a gold standard where we imagine all the gold in the world is money, and its quantity is approximately fixed".

This is cheating, as this precedes Nick's point 1. Imagine a world where the quantity of money is fixed. Like nations adopting Bitcoin as a currency. If monetary policy target would be to have constant monetary base, then we do not need to have Central Banks, there are no reserves - as Bitcoin is money. There is no CB and there is no monetary policy to speak of since there is only one think it "does" - it "sets" fixed money base and promises to "do nothing" in the future. Ever.

It is exactly when you want to keep options such as revaluations and devaluations etc., that it is for instance convenient to introduce local national currency. CB could for example promise to keep the amount of this currency in some ratio to Bitcoin and then break this promise if it needs to. CB may use capital controls or other instruments to manipulate outflow of Bitcoins from the country. And this is the point where (I think) Nick's example starts. If you want to have CB that keeps all these options, is it that much different from having CB that targets CPI basket of goods? And I agree that it actually is not fundamentally different. And that all these things are there to give Central Banks power to manipulate money stock (and thus price level) in their domains.

[Nick Rowe]

dwb: If I delete 6, we have a central bank that is price level path targeting, rather than inflation targeting. If I delete 5, I go from flexible inflation/price level targeting to strict inflation/price level targeting. There is only limited flexibility under inflation targeting. But there was also some flexibility under the old gold standard, due to transportation costs and transactions costs of gold, IIRC, the "gold points" had a small gap between buying and selling prices?)

[Bill Woolsey]

I see, like usual, I have too many typos.

Also, the spam detector is very difficult to manage.

(Maybe it is supposed to stop me!)

One final thought.

With a CPI target, if there is an increase in the demand for bling, the price of bling rises. Those buying the bling spend less on other goods, and the prices of those goods falls. There is no impact on the price level or the CPI. The central bank does nothing.

One slight problem occurs if the price of bling rises, but the prices of the other goods don't fall. This raises the CPI, and with price level targeting (especially tight targeting) the central bank contracts so that other prices are forced down.

With nominal GDP targeting, the increase in demand for bling raises the price and quantity of bling, raising nominal GDP. But whatever it is that those who were buying bling buy less of as a reduction in price and/or quantity. Even if the prices of those goods don't fall and it is rather their quanties that fall, there central bank does nothing. Total spending remains on target.

And that is a key reason why nominal GDP targeting is better than a gold standard. An increase in the demand for bling doesn't require a monetary contraction to deflate prices in response to an increase in the demand for bling.

[Nick Rowe]

wh10: "I think the transmission mechanism fundamentally changes when you change from a convertible regime to free floating. I think changes from quantity of reserves to interest rate on reserves."

But, we don't really have free floating. We have gone from (direct) convertibility into gold to (indirect) convertibility into the CPI basket.

[wh10]

My brain hurts:
[edited to embed link. I remember monetary theorist Kevin Dowd quoting exactly that line once during a seminar he was giving! NR]

[Dan Kervick]

I might be missing the point here, Nick, but it seems to me that you are just asking what is the difference between a regime in which the central bank targets the price of its currency relative to a single commodity, gold, and a regime in which it targets the price of its currency relative to some basket of commodities. But I think there was more to the gold standard than price targeting.

The salient fact about the gold standard is that the currency is legally redeemable in gold, and so the government actually has to own lots of gold. And gold is something the production of which the government does not fully control. It has a life beyond the government, and its mining, transportation and refinement are costly. The modern government does not typically manage consumer price stability by buying and storing lots of pork bellies and laptops, and offering to redeem its money for them. It relies primarily on its control of only one side of the exchange - the base money whose production at nearly zero cost is entirely within its discretionary power.

Are fiat currency units a claim on the nation's product of goods and services? Yes, in a certain weak sense, but only because the government's fiat money is generally accepted, and with a purchasing power that is left up to the seller. I can't take my US dollars to a pork belly bank where they are guaranteed to be redeemed in bacon. What the government does is to try to assure that the currency will continue to be accepted. It does this by enforcing various kinds of laws which require people to transact business in that currency. It imposes taxes for one thing, payable only in that currency, and requires banks to transact business among themselves in that currency.

If the government merely controlled the production of fiat money, but had no tools for making people accept it and use it, there would be no monetary policy at all. The prices and exchange mechanisms for commodities and services would move in a manner that was entirely unmoored form the government's alleged "money". In the gold standard era the government dependently piggybacks on the fact that gold was already generally acceptable as a medium of exchange, for historical reasons that had little to do with the government's own decisions. The government then attempted to peg its currency to the gold through redeemability promises and worked to manipulate the flow of gold itself though its economy - and even the global economy - so that gold prices relative to other commodities would be favorable. In the modern fiat money era, the government uses more direct tools for requiring people to employ its money as a medium of exchange, and for stabilizing prices of commodities in that currency.

[K]

Nick: "We have gone from (direct) convertibility into gold to (indirect) convertibility into the CPI basket."

You say that, but it can't be true. Lets say a supply shock causes a 50% drop in output. Now I, and the rest of Canadians, want to convert our money into consumption basket. Now, please! What means does the BoC have at its disposal to provide that? The answer, of course, is none. All they can deliver is Canadian dollars, which will inflate without massive (impossible) taxation support from the government. This is admittedly an extreme example, but it points to a basic structural fallacy. The government has only limited taxation means to deliver the consumption basket, and if people begin to expect to satisfy an excessive portion of their future consumption via debt/money assets, i.e. "indirect" consumption guarantees, the government will not have the ability to deliver on the guarantee. Step 8 fails because you did not replace the gold with consumption basket, but instead you replaced with a (limited)  taxation asset. You just waved your hands and pretended it didn't matter. It does matter.

[Mike Sproul]

OK. This is a little scary, since your post makes perfect sense to me. Just one question:

Do you think the money-issuing bank needs to hold enough assets to be able to buy back every dollar it has issued at par? I remember when you used to say that money-issuing banks didn't need assets, or at least they only needed assets equal to maybe 30% or so of the money they had issued.

[Nick Rowe]

Mike: "Do you think the money-issuing bank needs to hold enough assets to be able to buy back every dollar it has issued at par? I remember when you used to say that money-issuing banks didn't need assets, or at least they only needed assets equal to maybe 30% or so of the money they had issued."

Damn good question! One that's a little scary to me! Two thought's:

30% assets is enough unless there's a really big drop in demand for base money. (But what about the risk of runs??)

A bigger drop in demand for base, without enough assets, would maybe cause it to have to drop the peg to the CPI? (But it could borrow against its own future seigniorage revenue.)

All: Sorry I'm not responding to all comments. gotta do my taxes and write an exam.

[Tom]

Nick, I really like your attempt here, and it seems about as fair to me as it could be.
Like Ryan (quoting Mises), the first big change is 1>2, fractional reserves.

The instability of fiat money is more because of fractional reserves, than because of gold, gold+silver, or CPI basket.

Going back to a gold standard, while allowing fractional reserves, would not flush out the monetary instability.

The desire for fractional reserves is because of the deflation that would be caused by a "fixed" amount of money with increases in production, and such deflation reduces the growth rates. In order to get higher growth rates, "a little" money instability has been acceptable. (This is why I'm not quite a gold-bug.)

It might be that the injustice and instability of the various fiat moneys has gotten so bad, that reduced growth might be a price worth paying to get rid of the CB fiat system(s). This would be pushing the "all gold is money" view to become the dominant view. (Perhaps augmented by: Silver, Platinum, Palladium?) (I'm moving towards this side of the tradeoff.)

Finally, note that gov't deficit (over-)spending remains the big political issue, which no monetary change will solve alone. A gold standard makes it tougher on the gov't to do the overspending, and that seems a big and increasing benefit.

[Bill Woolsey]

K:

I find this shocking.

Wow.

If there is a 50% drop in output, then real incomes fall in half. With a stable price level, this requires nominal incomes to fall in half as well.

There is no guaratnee by the money issuer to provide sufficient consumer goods and services to keep real income from falling.

With real income falling in half, so does real tax revenue. With the real national debt unchanged, national bankruptcy is possible. This would involve all government debt and not just any money issued by the government.

With real income falling in half, the real demand for money will fall approximately in half. To keep the price level from rising, the quantity of money must fall in half. With the usual arrangements, the central bank must sell off half of its assets, generally government bonds. Again, if tax revenues falling in half means national default, this will be impossible.

With a zero percent gold standard, there is no difference, unless "Canada" is an important part of the world gold market, and the reduction in output doesn't apply to gold. Nominal income still falls in half. The Canadians still reduce their money holdings by half (presumably to import goods.) The Candadian central bank still has to cut its money issue by half by selling government bonds. And if Candada defaults because tax revenue falls in half, then it defaults on the money too.

With a 100% reserve gold standard, then even if the 50% drop in output results in default on the national debt, those holding money don't take a loss. Again, with Canada being small, there is no effect on the price level. Nominal income falls in half. Gold is exported to buy consumer goods. The quantity of gold money in Candada falls in half to reflect the 50% decrease in the demand to hold money caused by the 50% decrease in real and nominal income.

[Nick Rowe]

Dan: "The salient fact about the gold standard is that the currency is legally redeemable in gold, and so the government actually has to own lots of gold."

But, AFAIK, central banks typically had much less than 100% gold reserves. Plus, they could always borrow gold reserves if they needed to!

[BT]

This is progress. Things clearly evolved from the gold standard days to the modern context. The money multiplier is a good approximation of gold-standard bank behavior. But today things have become different enough to warrant a new description. And getting the description right is important if you want to build a Keen-style systems engineering model of the economy.

Two points:

1. The gold standard was fixed for many years at a time. Inflation targetting rarely hits 2% consistently. It's more like a big range of 0% - 10% that fluctuates quite a lot in most economies.

2. Bank notes were the main means of exchange between people for a long period of history. Credit transactions (between deposit accounts that are never withdrawn as cash) are the main means of exchange between people in the modern world - reserves just account for transactions between banks. The ratio of banknotes in circulation / total credit seems to have shrunk quite a lot over time. This change makes a difference.

[JP Koning]

"We have gone from (direct) convertibility into gold to (indirect) convertibility into the CPI basket."

I wouldn't describe it that way. Under a gold standard, the convertibility feature provided by central bank money - when it was exercised - was to be settled in gold by the central bank. The convertibility rate was some fixed quantity of gold. Everyone could directly go to the central bank and enjoy money's gold convertibility feature. Convertibility meant that in the secondary market, the public market for already-issued money, central bank money could purchase the same amount of gold for which it could be converted at the central bank.

Nowadays, upon exercising central bank money' convertibility feature, the redemption medium is bonds, not gold. You can redeem notes for bonds via open market operations. The redemption rate is adjusted to ensure that, in the secondary market for central bank money, its value relative to goods-in-general falls by 1-3% a year. Only a select few institutions can enjoy this redemption feature, but their participation is enough to ensure that central bank money falls at a 1-3% rate.

In sum, we haven't gone from gold conversion to CPI conversion. The modern convertibility feature provided by central banks doesn't allow for redemption of money into the CPI basket. It allows for conversion into bonds at some floating rate such that the secondary market for central bank money depreciates at 1-3%.

[Mike Sproul]

"But what about the risk of runs??"

Also a damn good question! As soon as assets drop to the point where the issuing bank is unable to buy back all its dollars at par, rational customers will run on the bank. So all banks need 100% asset backing. If they can squeeze a little extra from future seignorage revenue then that should have been counted as an asset in the first place

[123]

@Nick
@Bill Woolsey
@Scott Sumner

there is no fundamental theoretical difference.

All these ten banks are financial intermediaries with two targets. They all target expectations of some macroeconomic variable and they also manage the risk premium on those expectations. They all need equity so the targeting is credible.

They all differ in the type of the target, in the degree they allow risk premia to fluctuate, the need for equity is also different.

Number one has a target that is not very useful to customers, but it requires little equity and ensures that the risk premium on the targeted variable is always tiny. It is very easy to set up - witness the proliferation of the physically backed gold ETFs. I don't agree with Bill when he said that it offers unfair advantages to jewelers.

Number two has a target that is not very useful to customers, it requires more equity, and the control of the risk premia is not so good. It is much more successful because many customers like Bill Woolsey feel that the risk/return trade-off these banks offer is more attractive. On the other hand, many goldbugs say that number two is a sin and a fraud.

The challenge is to choose the most sensible macroeconomic target, for most large countries it is NGDPLT. It is also needed to pick the appropriate bounds for the risk premium, perhaps leaning against the Minsky cycle. Scott Sumner says the risk premium should be always zero, but then the infinite equity is needed. In practice, the equity of the central banks is a scarce resource.

[Nick Rowe]

BT: "This is progress. Things clearly evolved from the gold standard days to the modern context. The money multiplier is a good approximation of gold-standard bank behavior."

I think you might have missed the point of the post, which is to say that things have *not* (fundamentally) changed since gold-standard days (a point that surprised me). Also, see the bit in the first link to Edward Harrison's second post where he says:

"We have been living in a world predominated by floating exchange rates and currency non-convertibility for forty years now. Nevertheless, most of economics world seems to take a fixed exchange rate, Bretton Woods, or gold standard view of money and banking. In that world, as Warren Mosler quipped, bank lending is reserve constrained with the interest rate an endogenous variable via bank competition for reserves."

You might say this post is regress. ;-)

[Nick Rowe]

JP: I think I disagree.

Suppose initially in step 1 the central bank didn't actually own any gold, simply because it didn't own a strong enough safe, but gave anyone who wanted to redeem their notes in gold a sufficient amount of copper that they could sell the copper for the requisite amount of gold in the open market.

nemi: "Isn´t this a strange question?
Essentially, a fiat currency simply has more degrees of freedom than one that is backed by e.g. gold, i.e. the possible actions/strategies under a gold standard is (at least almost) a subset of those possible under a fiat currency."

Re-frame: targeting the price of gold and targeting the price of the CPI basket are just 2 out of the endless possible things you can do with a fiat currency.

[K]

Bill:

"I find this shocking.

Wow."

I assume you mean shockingly stupid? I'm not sure I totally get which part was the most stupid, but (possibly among other stupidities) perhaps it's this?:

"There is no guaratnee by the money issuer to provide sufficient consumer goods and services to keep real income from falling."

No. But inflation targeted government money/debt constitutes a consumption guarantee. If real incomes collapse, people will attempt to make up the shortfall by spending their consumption guarantees before everybody else does. If there are a lot of consumption guarantees (govt money/debt) outstanding, then that will cause inflation, i.e. the consumption guarantee will fail, much like in the story you tell. If there was explicit backing, then, as you say, the currency can't fail. But explicit backing is not possible under inflation targeting because the consumer basket is not a capital asset. In effect, nobody can guarantee any particular level of consumption. All that can be guaranteed is whatever capital assets can yield, which is a risky quantity.

I have great respect for your opinion, Bill, so I fear I'm still being stupid. I really don't get it.

[Nick Rowe]

K: (You are never stupid, BTW, but I think you are off on this one). If I own a $20 BoC note the BoC guarantees how many CPI baskets I can buy with that note (the guaranteed amount falls at 2% per year, and has a small error band).

[Alex]

I think that those interesting discussions need another format. I would suggest Nick to follow "Wilmott forum" in organizing discussions. It would be more convenient to participate in my view.

http://www.wilmott.com/index.cfm

[K]

Nick,

Thanks.

"If I own a $20 BoC note the BoC guarantees how many CPI baskets I can buy with that note"

Yes! Which is what I keep saying. But if Canadian output drops to say, zero, then it's going to be really hard for the BoC to live up to that guarantee, no? Suddenly that twenty is going to buy no baskets at all, right? Ultimately they'll need the tax collecting power of the Federal government *backed* by the productive output of the Canadian economy in order to guarantee their ability to make good on their promise. Yes they need the force of law. *And* they need the power of taxation.

If, on the other hand, the dollar was explicitly backed by gold, then, as Bill says, they could just deliver gold and then, to the extent that the exchange rate between gold and the consumption basket is unaltered, the consumption guarantee would be good. (Golden, in fact!)

I'm sure I'm still making no sense, cause people keep saying stuff to me that I already know as though it's obvious that I'm violating self evident truths. And I'm not seeing the contradiction. :-(

I hope at least Dan Kervick agrees with me, since I thought he was saying the same thing, but I'm not hopeful.

[genauer]

trying to post a comment again

[genauer]

@ Bill Woolsey
You basically spell out, that a gold guarantee establishes just "35 dollar = 1 ounce gold". I did and does not guarantee that you get a constant amount of other goods for it. That can go up or down, depending on supply and demand. People had high corn prices and hunger with gold pegged currencies.

So a Gold standard was and is not a CPI standard in our modern understanding, in both a good and a bad way.

[genauer]

@Chmee
this "the CB running the economy" comes from the anglo perspective that the CB is not just targeting low and stable inflation,
as it is the Bundesbank doctrine, and should be the ECB doctrine.

From that perspective the bank has basically just one measurement, the inflation, and basically one control parameter,
the interest rate, and controls the monetary policy.
Only 30 years ago, it was in fact the monetary supply, which was seen as the central lever, and not the rate.

[Ryan]

K - FWIW, what you're saying makes perfect sense to me. And it might make me unpopular for saying so, but it sounds a lot like the Ayn Rand criticism against inflation, if you're familiar with it.

[Oliver]

Interesting!

Promising to convert to something useless (gold) is both easier (less tied to real lives) and for the same reason less effective (less relevant to real lives) than promising to convert into something that people really need. How about a health standard? 1 Dollar promises to buy 1 day's worth of state of the art medical care.

Just a thought. Off to bed...

[Determinant]

So if CPI targeting isn't really that far from the Gold Standard, then that would explain why the Left sees monetary policy as a subset of the Right's agenda to promote inequality. The Gold Standard was the target of much Lefty criticism before the Great Depression sunk the concept.

It would also explain why Nick can't tolerate the Lefty position that Central Banks are irrelevant or tangential when the Left Agenda is political.

@Oliver:

Makes much less sense in countries with single-payer systems, that is anywhere except the US.

[dwb]

Re-frame: targeting the price of gold and targeting the price of the CPI basket are just 2 out of the endless possible things you can do with a fiat currency.

i agree completely. However, there is a world of difference between being short a physical (scarce) commodity with fixed supply (basically inelastic supply) and a financial one (whose supply is essentially infinite and can be suplied elastically). If you have to go out and physically procure and deliver it upon redemption you are faced with inelastic supply (been there, done that). Even if you fix a basket of goods, if you are forced to actually redeem or deliver those goods (yes, a *deliverable* basket of CPI) then the CB may be faced with inelastic supply. the difference is that even though the central bank targets the CPI (gold price, whatever) it does so by supplying financial bonds (claims) which it has (virtually) infinite supply of.

[Nick Rowe]

K: OK. I think I maybe get it. Let me ask a different but related question: the stock of money is a stock, and the flow of consumption is, urrrr, a flow. Just suppose everyone wanted to convert his base money immediately on CPI goods. It can't be done. You can't spend (say) 10% of GDP in 1 week. Interest rates would go very high, to try to persuade people to convert their base money into a stock of bonds, not into a flow of consumption. The central bank might go bust, because the market price of its bonds would fall as interest rates increased.

[wh10]

http://www.cnbc.com/id/46970524

[Nick Rowe]

Maybe, just maybe, the theoretical difference comes when we target prices that are sticky, as opposed to targeting prices that are perfectly flexible?? So it's at step 8 that the theory has to change?

[Nick Rowe]

Bill @8.32am. That is a very good comment, if a little hard to follow. I think you are right.

Short oversimplified version of Bill:

1. A world in which most of the stock of gold is used as money is theoretically very similar to a world in which the central bank targets the stock of paper money.

2. A world in which most gold is used for industrial purposes and paper is pegged to gold is theoretically very similar to inflation targeting.

[bond guy]

At what point in the above steps did monetary theory need to fundamentally change?

1. Start with a monetary system where the central bank's paper money (or an electronic equivalent) is 100% backed by gold reserves, and each note can be redeemed at a fixed price for gold.
...
3. Now take the limit of that system, as the percentage of gold reserves approaches 0%. There is indirect convertibility of paper money into gold. The central bank never actually buys or sells gold, but it buys or sells bonds to keep the equilibrium price of gold pegged, so that people can still swap their notes for gold in the open market at a fixed price.

Although some comments are focusing on the inability to buy or sell the CPI, the breakdown occurred earlier
in your logic. You leaped from a fixed price to an equilibrium price at step 3. That won't work in the real world (which theory should presumably reflect). Why: open market operations in the local market are unlikely to be enough to stabilise the price of a globally traded commodity in local currency terms.

As an example, suppose Canada had the idea to target the gold price in CAD terms without going into the gold market. Then assume the price of gold rises 10% in USD terms because of foreign buying. According to your scenario, the Bank of Canada would have to raise rates sufficiently to increase the value of CAD versus USD by 10%. That would presumably take a lot of rate hikes. Since this rate hike policy will be damaging to the Canadian economy, CAD could drop in value faster due to investor flight than would be made up by people positioning for interest rate differentials. If this policy damages the economy enough (e.g., policymakers being idiots destroys the incentive for fixed investment), CAD may never restore its original exchange rate.

Therefore, there is no reason to expect that open market operations could restore the CAD gold price on any reasonable interval, which makes a mockery of the assumption that the "equilibrium" price is fixed.

[JP Koning]

"Suppose initially in step 1 the central bank didn't actually own any gold, simply because it didn't own a strong enough safe, but gave anyone who wanted to redeem their notes in gold a sufficient amount of copper that they could sell the copper for the requisite amount of gold in the open market."

Ok, but I'm not sure why this is relevant. Then you have bank money convertibility which is settled with an amount of copper. The redemption amount is manipulated by the central bank to ensure that it's liabilities trade in the secondary market at a set ratio to the gold price.

That doesn't change the fact that the convertibility feature that modern central bank's provide to holders of their liabilities is settled with bonds, not CPI. The fact that people can sell these liabilities in the secondary market for a CPI basket is not convertibility. Convertibility is a feature that characterizes the financial relationship between a liability issuer and holder *in the primary market*. It allows the issuer/holder of an asset to destroy that issued asset rather than allow it to continue hot-potatoing around in the secondary market. In the case of central bank money, it is withdrawn/destroyed in the primary market by conversion into bonds, not in the secondary market by an exchange of CPI baskets. You don't convert money in the secondary market, you sell it.

[Nick Rowe]

JP: but under the gold exchange standard, AFAIK, most convertibility was in practice indirect. Central banks mostly used open market operations in bonds, rather than gold market operations. They had to, because they had fractional reserve banking. Plus, AFAIK, non-US (or earlier, non-UK) central banks had indirect convertibility into gold. You exchanged your local currencies for USD (or Sterling), which in turn was converted into gold.

bond-trader. Your argument seems to imply that the gold standard was impossible, with fractional gold reserves. Now sure, it was vulnerable to speculative attacks, but it did in fact exist.

[Determinant]

It just wasn't as robust as we thought it was, witness WWI and the tribulations of the 1920's where it was a curse.

[K]

Ryan: " FWIW..."

It's worth a lot. It means the number of people who apparently agree just doubled. I googled the Randian take on inflation and it sounds similar, except she seems to think that money should be a substitute for the consumption basket. I think it should be capital assets, or equivalenty, I think it should be a proxy for the intertemporal consumption basket, not just the current one.

Nick@7:12,

Agree with all of it. The only thing I'd add is that the CB can be made whole via taxation, assuming the losses aren't too great. But if the quantity of money and debt is too high, it may cause an impossible tax burden and the target will fail.

[JP Koning]

Nick, using your definition of convertibility, which I find somewhat peculiar, I agree with you. Convertibility to you seems to mean just selling in general.

Using my definition, can you see why I'd disagree with your point "most convertibility was in practice indirect."? To me, indirect equates to transactions in the secondary market. Direct is the primary market. In finance, convertibility by definition is not a mere sale, it occurs when the original issuer is forced to redeem by the holder, and that only happens in the primary market ie. directly. You can't have conversion in the secondary market. The phrase "convertibility was in practice indirect" is therefore nonsensical. The secondary market is where on-trading of already-created assets occurs, not conversion.

Anyways, ignoring all that, and using your wording now. Would you agree that indirect conversion of modern money into the CPI basket is achieved by direct conversion of central bank money into bonds via open market operations (and the threat thereof)? You said earlier "We have gone from (direct) convertibility into gold to (indirect) convertibility into the CPI basket." In my earlier posts, I was just pointing out that we've gone from (direct) convertibility into gold to (direct) convertibility into bonds.

[Mike Sproul]

JP:

What's the difference if one kind of dollar promises redemption for the CPI basket, and another promises redemption for the CPI basket or something of equal value, including bonds? Both meet the usual definition of "convertible".

[Nick Rowe]

JP: "Would you agree that indirect conversion of modern money into the CPI basket is achieved by direct conversion of central bank money into bonds via open market operations (and the threat thereof)?"

Yes.

BTW, I think my usage of the term "indirect convertibility" is standard. I didn't invent it.

[Nick Rowe]

My classes are now over, and exams haven't started yet, so I'm going to be away from the computer for a few days to take a needed break. Have fun.

[bond guy]

Nick Rowe:
bond-trader [sic]. Your argument seems to imply that the gold standard was impossible, with fractional gold reserves. Now sure, it was vulnerable to speculative attacks, but it did in fact exist.

The Gold Standard relied on the banks having currency and gold reserves to defend parities; your example had no reserves. Normally, speculators would act to stabilise gold parities on their own. But they would only do that if the (collective) central banks would act to defend those parities, and the reserve positions remained credible. Central banks did often have to launch coordinated moves to defend various parities of the weak links of the system (typically trade deficit countries). I re-read Eichengreen's "Golden Fetters" recently, and there is a long set of discussions about this point. Your assumption that you can progressive lower gold coverage ratios and still have a credible gold peg is thus suspect. (If a private actor could take out your gold reserves single-handedly, your gold peg would last roughly until lunch time.)

A country cannot peg a minimum value in an external currency without having reserves to back it up. (You can stop appreciation, as Switzerland is currently doing.) If nobody wants to own your currency (e.g., Zimbabwe), it's going to lose value in nominal terms versus external pairs regardless of your domestic interest rate policy. (Of course, even if you have reserves, currency pegs still don't last forever.)

Eventually, you can deflate your economy enough that purchasing power parity concerns should restore the original parity. But that process can take up to 20 years in practice (domestic prices are sticky to the downside, yadda yadda). Meanwhile, a thinly traded international commodity like gold could be rising in the foreign currency's terms (think Peak Gold), so a gold target could recede faster than you can deflate to keep up.

A regime in which a price which is pegged (within a window) each and every day (as per the Gold Standard) is not the same thing as a regime where the price which might drift back to its original level within 20 years.

[JP Koning]

Mike/Nick, then we just have different terminology. Your's might be standard in the monetary econ literature, but convertibility means something quite different in the financial world. In the end we're arriving at the same idea.

[JP Koning]

DWB: "the difference is that even though the central bank targets the CPI (gold price, whatever) it does so by supplying financial bonds (claims) which it has (virtually) infinite supply of."

It doesn't have an infinite supply of bonds. Just like classical central banks who had to buy gold in the secondary market, modern central banks must buy bonds in the secondary market along with everyone else. Sure, gold markets tend to be less broad and liquid than bond markets. But rather than a world of difference, I'd describe it a slightly different shade, and certainly not enough to contradict Nick's progression from 1-10.

Dan K: "The salient fact about the gold standard is that the currency is legally redeemable in gold, and so the government actually has to own lots of gold. And gold is something the production of which the government does not fully control...I can't take my US dollars to a pork belly bank where they are guaranteed to be redeemed in bacon. What the government does is to try to assure that the currency will continue to be accepted. It does this by enforcing various kinds of laws which require people to transact business in that currency...In the gold standard era the government dependently piggybacks on the fact that gold was already generally acceptable as a medium of exchange, for historical reasons that had little to do with the government's own decisions. "

The salient fact about the modern standard is that the currency is redeemable in bonds, and so the central bank has to own a lot of bonds. Bond production is something which the central bank has no control over. The only difference therefore between the gold standard and the modern bond standard is we've switched redemption medium and allowed the redemption price to float. That's why the theory has remained constant, though details have changed, as Nick points out.

Just like a gold standard bank, a modern central bank doesn't need enforcement of laws in order for it's currency to be accepted. The fact that one offers convertibility into gold and another into a floating quantity of bonds (not porkbellies) is sufficient reason for both to remain credible.

While in the gold standard era central banks could piggyback on the fact that gold was already generally acceptable and had value, modern central banks can piggyback on the fact that bonds are already generally acceptable and have value. Nick's progression from 1-10 stands as far as I can see.

[y]

It's possible to think about contemporary fiat money systems in the same way that one thinks about gold-standard systems.

This kind of approach misses the point however that fiat money makes it possible to do things in a very different way.

In a fiat system the government/CB essentially has a magic goldmine that can produce an infinite amount of gold.

If that's no different to a gold standard system with a limited supply of gold then I don't know what is.

In this magic goldmine system state borrowing isn't really necessary.

Financing government deficit spending through money creation (or CB "monetising" government debt) isn't necessarily more inflationary than deficit spending funded by bond issuance.

This is because interest rates can be determined by paying interest on reserves rather than by selling/buying bonds. This means the overall quantity of reserves is not important - even if the government chooses to pursue a policy of controlling inflation through interest rates.

Piling up excess reserves does not necessarily make it easier for banks to lend, nor does it increase the availability of currency (cash) for withdrawal by depositors (CB always makes cash available anyway).

Any level interest rate can be sustained through the paying of interest on reserves, regardless of the overall quantity of excess reserves in the system.
Alternatively the interest rate can be allowed to fall to zero.

With a fiat money system taxation can be viewed as a means of regulating the currency's value rather than as a means of raising revenue.

Thus it might be more sensible to control inflation mainly through fiscal policy, and to keep interest rates stable (thereby avoiding the sledge-hammer, one-size fits all approach of monetary policy).

Fiat money also implies that high government deficits are not necessarily unsustainable, so long as government spending is not overly inflationary, wasteful, inefficient or harmful in some other way.
Government deficit spending always has the potential to be inflationary, but whether it is financed through bond issuance or "money-printing" is not fundamentally important.

For these reasons magic goldmine is very different to ordinary goldmine.

[123]

"Short oversimplified version of Bill:

1. A world in which most of the stock of gold is used as money is theoretically very similar to a world in which the central bank targets the stock of paper money.

2. A world in which most gold is used for industrial purposes and paper is pegged to gold is theoretically very similar to inflation targeting."

Both types of central banking from the #1 above have the advantage of requiring little equity and having a very low level of risk. However, the services they provide fail the market test, so we do not observe #1 often in practice (Volcker tried to move into the direction of #1, but this was temporary). #1 have the current day private sector equivalents of physical gold ETFs and closed-end currency mutual funds. Their impact is tiny.

#2 is much more useful and popular, having better risk/reward ratios, Rothbardian moralizing notwithstanding.

[y]

* I should add that owning a government bond does not necessarily restrain or 'hold back' potential spending by bond owners, as bonds can be sold quickly and easily.

* piling up excess reserves does not make it easier for banks to extend credit as reserves are necessarily always available under the current system. What's important is the price, i.e. the interest rate. This can be set by paying interest on reserves rather than through OMOs.
The efficacy of controlling inflation through interest rates should also be questioned - does it actually deliver optimal outcomes, or even work that well?
Zero interest rates (or very low rates) combined with a different approach to fiscal policy and regulation might actually deliver better outcomes along with the desired level of (low) inflation. Very low rates are not necessarily stimulatory *in the way that is usually assumed*, given that they reduce interest income.

[y]

Let's say we're on a 'CPI standard'.

The main difference I would say with the gold standard is that, as I said above, the government is not really revenue-constrained under a fiat money system.

As such the only real financial constraint it faces is inflation. This means government has to ensure that its spending doesn't lead to inflation, but doesn't have to worry about 'going bust' or 'running out of money', in the traditional sense. Similarly, it needs to be careful that it's spending is not overly wasteful or inefficient, as this could lead either to inflation or to other structural and behavioural problems in the long term. It also needs to be able to carry out an appropriate taxation policy so as to be able to regulate the basic demand for/ value of the currency.

The central bank can also try to ensure the inflation target is met by setting interest rates correctly. It has the option to set this rate by paying interest on reserves rather than through open market operations.

As such government debt is not necessary, either for the government to deficit spend, or for the CB to set interest rates.

Either we can have a system in which the government issues new money directly, one in which the CB buys government debt as it seems fit, or one in which it buys all government debt and simply pays IOR.

In each case the money needed is simply dragged out of the magic infinite gold mine.

There is no need to worry about bankruptcy in the traditional sense because the magic goldmine never runs out of new money.

So policy, both fiscal and monetary, can be properly conducted to achieve optimum outcomes (max output/ low inflation) without having to go through periodic bouts of deficit/debt hysteria. Government and central bank can focus on real problems rather than having nightmares about phantom bond vigilantes.

This is potentially quite different to a perennially dysfunctional gold standard system, in which bog-standard gold mines consistently fail to produce the right quantity of gold at any given point in time.

Magical infinite gold mines are much better, if they're used properly. They mustn't be abused however, or they could disappear in a puff of smoke!

[Determinant]

Nick:

bond guy is more correct than he suspects. His posited case is not theoretical, that is precisely what happened to the USD/CAD exchange rate in the late 1940's. Immediately after WWII Canada and the United States were the only two industrialized nations that dealt with each other on an equal basis in money matters; Canada was not in debt to the US because we never participated in Lend-Lease, there had been political conferences during the war between Mackenzie-King and Roosevelt on this very point. Mackenzie-King convinced Roosevelt that placing a debt yoke on Canada was counter-productive and useless and would only injure the American reputation and their interests in Canada.

So the USD/CAD rate, uniquely, was determined according to trade flows alone, except that like everyone else it was fixed by Bank of Canada fiat until 1950.

In the late 1940's there was market pressure to devalue the CAD against the USD, of course this led to repeated intervention and threatened exhaustion of Canada's reserves of USD and gold. Of course we could have implemented capital controls, but we wanted a free market in investment more than we wanted a stable exchange rate. We could have changed monetary policy but we wanted an accommodating, independent monetary policy and nobody in this country wanted to choke off the post-war boom, not after living through the Depression and years of war. We wanted to enjoy the good times, not end them over a technical exchange rate.

Early on we encountered the Iron Triangle of capital controls, fixed exchange rates and independent monetary policy. We chose free capital flows and independent monetary policy and therefore had to float. 1950-62 was marked by exchange stability and economic prosperity so the governing elite in this country learned that fixed exchange rates and/or a gold standard was undesirable and unnecessary.

[JP Koning]

"This kind of approach misses the point however that fiat money makes it possible to do things in a very different way"

Y, you can still use Nick's progression-style of getting his point across and avoid missing the point you are trying to make.

Take a gold standard in which the convertibility of central bank money is set at a certain quantity of gold ounces. Say the central bank offers the government an unlimited line of credit. As the government spends through its central bank account, the public bring central bank money back to the central bank to be redeemed for gold as they don't need that quantity of cash on a day-to-day basis. The bank's gold reserves start to diminish and the credibility of the peg is drawn into question. In order to protect reserves and defend the gold peg, the government raises taxes, drains excess cash, and sends it to the central bank to be destroyed. The race for gold is dissipated and the peg to gold remains.

Now take a modern central bank that provides unlimited lines of credit for government. All that's changed is the redemption media is now bonds via open market operations, not gold. And the peg isn't fixed, it floats so as to force central bank money to fall in value at 1-3% or so a year.

Obviously modern central banks don't offer unlimited lines of credit. So while it is interesting to consider the possibility, Nick's way of explaining things resembles reality.

[Lord]

While we can speak of 100% redeemable currency, that does not mean all debts can be paid. Someone that owns then sells an asset financing the sale creates debt but no base money and someone that pays off their debt destroys it without affecting base money. This can create temporary cycles as lending standards relax then tighten even if money is fully convertible and debts that looked payable when created can fail to be so over the fullness of time. Nor would debts created in such manner necessarily raise interest rates or their destruction lower them since they are in soft money or income flows. A 100% reserve turns banks into depositories but cannot save the economy.