The Quantity Theory and Neutrality of Money when money is endogenous

"If the central bank permanently doubles the stock of (base) money: all nominal variables will double [that's the Quantity Theory of Money]; all real variables will stay the same [that's the Neutrality of Money]."

The Quantity Theory of Money and the Neutrality of Money go together. It is very hard to have one without the other. They are only exactly true under very strict conditions (see e.g. Patinkin's "Money, Interest and Prices"), which I will ignore here.

As stated above, the QT and NM implicitly assume that (base) money is exogenous. But it is easy to define versions of QT and NM that are applicable when money is endogenous.

For example, suppose the central bank fixes the exchange rate, which means it must allow the stock of money to adjust endogenously.

"If the central bank permanently doubles the price of foreign exchange: all nominal variables will double [that's the Quantity Theory of Money]; all real variables will stay the same [that's the Neutrality of Money]."

Or, suppose the central bank targets NGDP, which also means it must allow the stock of money to adjust endogenously.

"If the central bank permanently doubles the target level of NGDP: all nominal variables will double [that's the Quantity Theory of Money]; all real variables will stay the same [that's the Neutrality of Money]."

And so on. A doubling of the price of foreign exchange (or the NGDP target, or whatever nominal variable the central bank picks), will cause the stock of money to double, along with all other nominal variables. The stock of money is now endogenous, and the exchange rate (or NGDP) is now exogenous.

Whether true or false, the Quantity Theory of Money and the Neutrality of Money, are equally applicable in a world where the stock of money is determined endogenously. The same underlying idea of the classical dichotomy, that nominal variables (those with $ in the units) and real variables (those without $ in the units) are independent of each other (under some conditions), is equally applicable regardless of what is assumed exogenous.

But all this is just comparative statics, comparing one equilibrium time-path for nominal variables with another. The question of the stability of that equilibrium time-path is another question. For example, I think that stability would be much greater under NGDP level-path targeting than under base money stock targeting.

(Just an aside, for JW Mason.)

Comments

[Lorenzo from Oz]

Nicely clarifying, thanks.

[CMA]

"They are only exactly true under very strict conditions (see e.g. Patinkin's "Money, Interest and Prices"), which I will ignore here."

Couldnt you also say the s and p 500 is neutral too then? Its not exactly true but true enough. Everything is neutral.

The MB in the US from 1985 to 2005 increased about five fold while CPI less than doubled.

[Mike Sproul]

Problem:

If the central bank has issued $100 of base money, and holds assets worth $90 oz of silver, but fixes the exchange rate at $1=1 oz., then arbitragers will short dollars, customers will return dollars to the central bank and demand assets worth 1 oz, and the central bank will run out of assets after the first $90 are redeemed. At this point the central bank is broke and can't defend any exchange rate other than zero. Meanwhile the arbitragers get rich.

The only way to defend any exchange rate is to have 100% (or more) assets backing the central bank's money.

[Nick Rowe]

Lorenzo: thanks! Yep, I was aiming at a simple clarifying post.

CMA: Yep. If the central bank targets the level of the S&P500, which is a nominal variable, then the QT/NM says that a permanent doubling of the S&P500 target level will double all other nominal variables, leaving real variables unchanged.

"The MB in the US from 1985 to 2005 increased about five fold while CPI less than doubled."

Yep.

1. But the QT tells us what happens to the price level ***compared to what would have happen otherwise***, not compared to what it was in 1985. We do not observe the counterfactual. Other exogenous things were not (presumably) equal.

2. Plus, it is not just the *level* of the money supply that matters for the price level, it's the *growth rate* too. Compare two otherwise identical economies, with the same level of the base money supply at time t, but economy one has a higher growth rate of M than economy two. Since economy two has a higher inflation rate than economy one, and since the demand for base money is a negative function of the inflation rate (assuming base money pays 0% interest, which is correct for currency), then economy two will have a higher price level at time t than economy one, even though the level of the base is the same.

Mike: "The only way to defend any exchange rate is to have 100% (or more) assets backing the central bank's money."

That may or may not be true. But I would say its truth is independent of the truth of the QT/NM.

[Min]

Mike Sproul: "If the central bank has issued $100 of base money, and holds assets worth $90 oz of silver, but fixes the exchange rate at $1=1 oz., then arbitragers will short dollars, customers will return dollars to the central bank and demand assets worth 1 oz, and the central bank will run out of assets after the first $90 are redeemed."

I had trouble following that, so, con permiso, I have changed silver to yen. Doing so yields this:

If the central bank has issued $100 of base money, and holds ¥9,000, but fixes the exchange rate at $1=¥100, then arbitrageurs will buy yen and sell dollars, customers will return dollars to the central bank and demand ¥100 per dollar, and the central bank will run out of yen after the first ¥9,000 are redeemed.

Wait a second. Since when can customers demand yen from a central bank that issues dollars?

Mike Sproul: "The only way to defend any exchange rate is to have 100% (or more) assets backing the central bank's money."

Assuming that that is so, can't the CB retire $10 of base money to reach parity? The Lord giveth and the Lord taketh away.

[Min]

"If the central bank permanently doubles {nominal variable, N} all nominal variables will double [that's the Quantity Theory of Money]; all real variables will stay the same [that's the Neutrality of Money]."

OK. Suppose that the changes in nominal and real variables occur instantaneously (or as nearly instantaneously as possible). How does that happen? Mechanisms, please. Pick an N, any N. Thanks. :)

[Mike Sproul]

Min:

"Since when can customers demand yen from a central bank that issues dollars?"

Since Nick said this:

"suppose the central bank fixes the exchange rate, which means it must allow the stock of money to adjust endogenously."

"can't the CB retire $10 of base money to reach parity?"

In order to retire that $10, it must pay out 10 oz of its assets (or 100 yen). So while it used to have 90 oz of stuff backing $100, it now has 80 oz of stuff backing $90. That's 89% backing, even worse than the 90% it had before.

[Nick Rowe]

Min: You are talking about the stability experiment. Let N be M. Central bank doubles M. At the existing P there is an excess supply of M, so people try to get rid of that excess M by buying goods, which creates an excess demand for goods, which increases P.

[Mike Sproul]

Nick:

"That may or may not be true. But I would say its truth is independent of the truth of the QT/NM."

If it is true, then money is valued on finance-based principles, just like stocks and bonds, and the QT/NM go out the window.

[Nick Rowe]

Mike: suppose there's a 2 for 1 stock split. Only the stock is money. The helicopter delivers the new stocks, in proportion to existing stocks, just like in Patinkin. The only thing that matters is *how big* a difference does it make whether you use helicopters or OMOs.

[Majromax]

@Mike Sproul:

Your situation is impossible. As Nick noted and you quoted yourself, "suppose the central bank fixes the exchange rate, which means it must allow the stock of money to adjust endogenously."

That means that we can't a priori assume that the Central Bank has issued blank dollars against blank yen, because the CB *can't* unilaterally issue currency whilst defending its rate. You've essentially shown that indeed, the CB must allow the stock of money to adjust endogenously.

Your scenario is then further flawed because you're assuming that the CB can only alter the stock of money through redemption or issuance of currency for its backing. That's not true, as even exchange-rate-targetting CBs retain the power of the printing press. That means that the CB can helicopter money in or out as it sees fit. Should one find itself in your situation, then the obvious solution is to helicopter-lift $10 of base money away from the populace.

This, of course, would be politically unpopular.

[Mike Sproul]

Nick:

Suppose that before your helicopter drop, the central bank had 100 oz of assets backing $100. After the drop, the central bank will have 100 oz of assets backing $200. The backing theory says that each dollar will be worth half as much, since there is twice as much money for the same backing (just like a stock split). The quantity theory also says each dollar will be worth half as much, because there is twice as much money chasing the same goods in the market. (But the quantity theory ignores the central bank's assets.)

But if the new $100 was issued through open market operations, then after the OMO the central bank would have 200 oz of assets backing $200, so each dollar is still worth 1 oz. This is like a NEW ISSUE of stock, and is unlike a stock split. The quantity theory says each dollar would be worth half as much, in spite of the fact that each dollar is backed by 1 oz worth of stuff. Once again the quantity theory ignores the assets of the central bank.

Meanwhile, we have to consider the law of reflux. With the helicopter drop, real balances held by the public are unchanged, so there will be no reflux of unwanted dollars to the central bank. But with the OMO, the extra $100 is presumably not wanted in the circulation, and would reflux to the central bank, leaving the quantity of money unchanged. A complication: If the central bank refused to allow the unwanted $100 to reflux, then the central bank has effectively defaulted, and the effect would be the same as if the extra 100 oz worth of assets had been lost---there would then be inflation.

[Mike Sproul]

Majromax:
"You've essentially shown that indeed, the CB must allow the stock of money to adjust endogenously."

Which is to say, I've explained the law of reflux.

Yes, the CB can helicopter money, in which case the backing theory says there will be inflation, because there will be less backing per dollar.

Historically, governments (not central banks) have helicopter-lifted money from the populace by collecting money in taxes and then burning it. It was politically unpopular, but mostly because it caused money shortages.

[Frances Coppola]

Mike,

"The extra $100 is presumably unwanted in the circulation and would reflux to the central bank".

Sigh. The extra £100 remains in the central bank anyway. It is simply expansion of central bank liabilties. The monetary base never leaves the books of the central bank. The theory of reflux is fiction based upon a complete misunderstanding of how the monetary system works.

We can discuss to what extent the $100 actually circulates in the economy, if you like. Really that depends on the marginal propensity to consume of its recipients. You assume that because an OMO exchanges one safe asset for another, more liquid one, that therefore the recipients don't want to receive cash. Really? If they didn't want the cash they wouldn't sell. Indeed one of the mysteries of QE - which resembles a very large OMO program , though there are significant differences - is why investors whose marginal propensity to consume is approximately zero are willing to participate at all.

[Market Fiscalist]

Mike,

I have a question for you.

You say "Yes, the CB can helicopter money, in which case the backing theory says there will be inflation, because there will be less backing per dollar." which I agree with. Market Monetarists believe that inflation can be induced by swapping new money for other assets via OMOs. I think they believe this based purely upon the QTM (OMO changes M so changes in P must follow).

In my view OMO works primarily by changing the interest rate paid on bonds and so affecting inter-temporal spending decisions.

What role do you see interest rates having in the Backing Theory ? Do you think the monetary authorities can induce inflation by reducing the interest rates it pays on bonds (and encouraging more spending in the present) even though whenever the CB uses IR policy the entire money supply remains fully backed. If not why not ?

[Scott Sumner]

Nick, I certainly agree with all the economic analysis, but I don't think your definitions of money neutrality and the QTM are standard. You have defined money neutrality correctly. If the central bank increases the money supply by X%, that action will eventually cause all nominal aggregates to increase by x% relative to where they'd be without the increase. Real variables are unaffected. That's money neutrality.

On the other hand, the QTM is usually defined as making a much more ambitious claim, that actual real world inflation (or NGDP) is mostly caused by changes in the money supply, not money demand. Or at least that's the way I always interpreted it. For instance, you often see textbooks say that the QTM "assumes" V is stable. But money neutrality is true regardless of how unstable V is.

This is nitpicking. As I said, I think your economic analysis is exactly right. And I suppose you can find definitions of the QTM that fit any interpretation. Heck, I've seen textbooks confuse it with MV=PY!

[Scott Sumner]

I should have said "money is neutral in the long run" regardless of how unstable V is.

[Min]

Thanks, Nick. :)

What I had in mind was this. The claim is made that under very general conditions, change A will eventually lead to change B (which may be a set of changes). The question is under what special conditions does change A lead to change B very rapidly. The answer to that question may shed light on the dynamics of the general situation. Also, you can make the assumption that those special conditions will apply at some specific time in the future, and see what happens as that time becomes more remote. If the remote special conditions also cause change B to occur rapidly when change A is made, that is an argument against change A eventually causing change B, because if it is going to happen then, why isn't it happening now?

[Min]

BTW, if Mike Sproul is right that for $1 = ¥100, there must be 100 times as many yen as dollars, then the CB that issues dollars will have a hard time maintaining that exchange rate without the cooperation of the CB that issues yen. OTOH, if they agree between themselves on that exchange rate, what is the problem?

[Nick Rowe]

Scott: I think that different people do mean different things by "QTM". (Yes, some even call "MV=PY" the QTM, whereas I would say that MV=PY is just one of many ways of explaining QTM.) Milton Friedman, for example, seemed to understand "QTM" as a fairly broad approach to monetary theory. I tend to use use "QTM" in the narrow Patinkinesque sense. (Had Patinkin as a prof, briefly, and we also worked through Money interest and Prices.) And we always understood it in a counterfactual sense, so it would be true even if shocks to money demand were bigger than shocks to money supply. It's not so much that QTM assumes V is "stable", but that V is exogenous wrt the level of M.

Min: "The question is under what special conditions does change A lead to change B very rapidly."

The standard answer is that it depends (mostly) on how sticky prices are, and on how quickly expectations adjust.

[Mike Sproul]

Frances:

It is irrelevant whether $100 is issued in the form of paper dollars or in the form of deposit dollars. Both kinds of dollars show up on the liability side of the fed's balance sheet, and both are backed by the fed's assets. Only on quantity theory principles is the value of either kind of dollar determined by how fast they circulate, or if they circulate at all. On backing theory principles, the value of either kind of dollar is determined by how much backing the issuer holds against them---(which is just like stocks and bonds).

Reflux works for both kinds of dollars too. They are only issued if they are wanted. If not, the holder of either kind of dollar returns them to the fed in exchange for the fed's bonds or other assets.

[Mike Sproul]

Market Fiscalist:

I'd summarize my view of interest rates with an example: Say the market R=5%. If the Fed starts lending at 4%, then for every $100 lent, the fed loses $1, and we get inflation because there is less backing per dollar. Of course, with this low rate, loan demand explodes, and the fed's losses can become huge.

If the fed instead started lending at 6%, then in a competitive banking world, nobody will borrow from the fed, and other banks will fill the void.

[Too Much Fed]

Mike Sproul said: "Reflux works for both kinds of dollars too. They are only issued if they are wanted. If not, the holder of either kind of dollar returns them to the fed in exchange for the fed's bonds or other assets."

Mike S. and Frances, doesn't reflux apply to currency, but may or may not apply to central bank reserves? With QE, the commercial banks can't exchange the excess central bank reserves for bonds because the fed won't allow it?

[Too Much Fed]

Mike Sproul said: "Suppose that before your helicopter drop, the central bank had 100 oz of assets backing $100. After the drop, the central bank will have 100 oz of assets backing $200."

Nick, notice how Mike S. described his version of "helicopter drop".

[Market Fiscalist]

Mike,

You say: "Say the market R=5%. If the Fed starts lending at 4%, then for every $100 lent, the fed loses $1, and we get inflation because there is less backing per dollar"

That's a good answer. But what does it mean for the market rate to be 5% ? If the CB intervenes in the bond market isn't it causing a change in the market rate ? Plus: If the CB lowers rates, it increases the value of its existing holdings - shouldn't that cause deflation if the backing theory is true ?

[Lorenzo from Oz]

Mike Sproul: How would you describe the gold/silver/specie standard as working when the cover ratio was less than 100%? Surely the point was that as long as there was confidence that any given note would be redeemed, then notes did not need 100% backing.

It is not hard to see how notes on a gold/silver/specie standard was set by the value of gold/silver/specie, working on a supply and demand basis for the gold/silver/specie. So, under the classical gold standard of 1873-1914, assuming relatively constant demand for gold, you got mild deflation when output was increasing faster than gold supplies--1873-1896--and mild deflation when gold supplies were increasing faster than output--1896-1914. And really severe deflations when demand for gold suddenly shot up--1920-21, 1929-32. The value of money had nothing to do with banknotes as a medium of exchange, just the value of the medium of account (gold).

Note, this was a different question that of the supply/demand for the medium of exchange ("liquidity"). When the BoE was required to have 100% cover ratio (as per Peel's 1844 Bank Act), then when there was a sharp rise in demand for BoE banknotes due to some "liquidity crisis", the Chancellor of the Exchequer just sent the Governor of the BoE a letter releasing the BoE from the 100% cover ratio constraint. (There is no basis in English law whatsoever for executive suspension of a statutory provision, but folks apparently just let it go through to the keeper.) The backing did not change, indeed it "shrank" relative to the note issue, but liquidity would be restored.

But, under the gold/silver/specie standard, the medium of account as not "backed" by anything. It was just itself and its value was set by supply and demand for it.

Change the medium of account to Central Bank base money. Why is its value not set by supply and demand for it?

[Mike Sproul]

Too Much Fed:

The fed might refuse to allow reflux of either its paper money or its deposit money. Either way, it amounts to a default by the fed. This is tantamount to a loss of backing, and can be inflationary. But there are qualifications: Maybe the fed will resume reflux in 1 year, or maybe it will pay interest on its paper money or on its reserves until reflux resumes. In that case the effect of plugging a few reflux channels can be negligible

[Mike Sproul]

Market fiscalist:

I don't have a good answer for that. It seems too weird that a monopolistic central bank could gain from manipulating interest rates, but but I'll have to think about it some more.

[Mike Sproul]

Lorenzo:

If cover is truly .9 oz per dollar, then each dollar must be worth .9 oz. But "cover" is ambiguous. What if the government stood ready to bail out the central bank? In that case, there can effectively be 100% coverage, even if the central bank's books show only 90% coverage.

"Change the medium of account to Central Bank base money. Why is its value not set by supply and demand for it?"

Suppose the paper peso had no backing at all, just a limited supply of pesos plus some liquidity demand to give the peso value. The Mexican central bank would have originally issued each peso for (say) 1 oz of silver, but if that supply/demand view is right, the central bank doesn't need the silver. The silver is a free lunch to the central bank. If rival moneys, like the dollar, started to be used in Mexico, then demand for pesos would fall, and the peso would lose value, and the central bank's free lunch would fall, while the Fed grabs the free lunch for itself. One wonders how the peso could retain any value in the face of rival moneys.

On the other hand, rival moneys present no problem for the backing theory. There's no free lunch for anyone, and every currency is worth its backing.

[CMA]

If money is neutral why worry about monetary policy at all? Just issue a certain stock of money and forget about it right?

[Jacques René Giguère]

Lorenzo:"(There is no basis in English law whatsoever for executive suspension of a statutory provision, but folks apparently just let it go through to the keeper."
During the debate on the recharter of the Bank, Peel said something about ,IIR approximately,"patriotic men doing what is right for the country." This is legislator's intent and at the time carried weight (more than today). He was merely rephrasing the old roman principle of "salus populi suprema lex esto".

[Nick Rowe]

CMA: "If money is neutral why worry about monetary policy at all? Just issue a certain stock of money and forget about it right?"

That is a classic exam question. And the classic answer goes:

1. We need to distinguish neutrality from superneutrality: even if the level of M and P don't matter for anything real, the *growth rate* of M and P (inflation) might have real consequences (shoe-leather from higher V, costs of changing prices, relative price distortions, confused accountants, etc.).

2. Money might be long run neutral but short run non-neutral, because of sticky prices. So fluctuations in M (or fluctuations in MV) will cause undesirable fluctuations in Y (business cycles).

[Nick Rowe]

Lorenzo: my explication of Mike's position (Mike may or may not agree):

For some particular type of money M (Bank of Canada dollars): Put M/P on the horizontal axis, and (R-Rm) on the vertical axis, where Rm is the rate of return on holding M (the interest paid on holding money, plus capital gains from deflation) and R is the rate of return on other assets, so that R-Rm is the opportunity cost of holding money rather than some other asset.

If the Bank of Canada has some sort of de facto monopoly, either legal, or due to network externalities, then we get a downward-sloping demand curve for M. And the Bank of Canada can pick a point on that demand curve and earn the rectangle earn monopoly profits = (R-Rm).M/P That's the standard model.

For Mike's case, assume instead the demand curve is perfectly elastic (horizontal) at 0, because the issuer of M has no monopoly power, and is like a perfectly competitive firm. It must pay Rm=R to get people to hold its money, and earns no profits, and it must have 100% backing assets so it can pay Rm=R. The economy is always in a liquidity trap, and money is valued just like any financial asset. It's still demand and supply, but the demand curve is horizontal.

At the other extreme, where the demand curve is vertical, we get the very old monetarist model where V is independent of R-Rm. Still demand and supply, but the supply curve is vertical.

To my mind, all the empirical evidence points us to the first case. Zimbabwean dollars paid a highly negative Rm (due to inflation) but people still held them, so we can rule out the second case. But when Rm got very low, M/P fell to zero, so we can rule out the third case too.

[CMA]

"2. Money might be long run neutral but short run non-neutral, because of sticky prices. So fluctuations in M (or fluctuations in MV) will cause undesirable fluctuations in Y (business cycles)."

If money is "positive" on real measures in short run it would need to be negative at other times to offset and make it neutral in long run. Does that make sense? It seems as if money can be negative, neutral or positive depending on factors like structure of economy and and monetary system. Sticky prices may not be the only factor affecting money neutrality.

[Nick Rowe]

CMA: reading between your lines, here is a very simple early New keynesian model which has the properties you describe:

1. M+V=P+Y (that's MV=PY in logs)

2. P=E(M+V)-Y* where Y* is potential output, and E(M+V) is the previous period's expectation of (M+V), and V is an exogenous stochastic variable.

The solution is:

3. Y=Y*+(M+V)-E(M+V) Expected changes in M are neutral, but unexpected changes in M (or V) have real effects. And if expectations are correct on average in the "long run", then monetary policy has no effect on average Y in the long run, but a wise monetary policy, that offsets unexpected changes in V, will reduce the variance in Y.

[Philippe]

so in other words money isn't neutral.

[Nick Rowe]

Philippe: Instead of trolling, maybe read my comment immediately above yours, and try to understand it. Yes/no answers are sometimes not the most useful.

[CMA]

"Expected changes in M are neutral" in long run.

Im not quite sure this is true either though. If in short run expected changes in money are positive then it implies money is negative over other periods to be neutral in long run. Why would that be the case? Why would money be negative over periods outside the short term to offset short term positivity? Im pretty sure the fact that money isnt neutral in short run may strongly indicate that in long run it isnt either.

[Nick Rowe]

CMA: it will be exactly true in the little model I wrote down above. If M is less than E(M) (holding V=E(V)), then Y will be less than Y*. The unexpected fall in M causes a drop in output, a recession in other words. But if E(M+V)=(M+V) on average, then Y=Y* on average too, regardless of M.

But it is easy to change the model so it won't be exactly true. Just make the model non-linear, for example.

[Market Fiscalist]

Mike,

On "It seems too weird that a monopolistic central bank could gain from manipulating interest rates"

Perhaps the gains would be seen as only short-term ? The interest rate will eventually return to its true market rate and the value of the CB assets will also fall. In fact if the CB has been buying assets at above their long-term value (at low interest rates) they may take a capital loss when the market recovers. This could explain why the CB lowering interest rates often leads to inflation - in a way consistent with the BT.

[Min]

Here is a left-field thought experiment. Suppose that, for reasons unknown, the Bank of Canada declared that it would exchange $1 for 1 Bitcoin. The bank would be targeting the nominal value of the Bitcoin. Take a Bitcoin to the bank, get $1. That would provide backing for the Bitcoin, which it currently does not have. It would also provide a floor for the Bitcoin, in case it crashed.

Currently the exchange rate for 1 Bitcoin is around $640 Canadian. But it is quite conceivable that the Bitcoin could crash someday. (Moi, I think that it is likely, but who knows.) In that case people might well exchange bitcoins for Canadian dollars on a one to one basis. The Bank of Canada will have reached its target. Do we think that other nominal variables will have dropped to 1/640 of their current prices?

[Scott Sumner]

Nick, That all sounds right. But I do think my interpretation is pretty widespread. For instance, almost all mainstream macroeconomists accept that money is neutral in the long run, but many are critical of the QTM precisely because they believe V is unstable. And again, most textbooks talk about stable V as an assumption behind the QTM. They don't generally present it merely as a counterfactual. Patinkin's version is the version that, of course, is much much likely to be true. (I had Patinkin as a professor too, but unlike you I don't recall much of what he had to say.)

Oddly, early "cointegration" tests that "refuted" the QTM made the opposite mistake. They correctly (from my point of view) understood the QTM to be a model that predicted stable V, but they wrongly assumed that "stable V" meant "absolutely constant V." By the criterion of the cointegrationists, all economic models would be "false."

[Nick Rowe]

Scott: I think we are on the same page here. The only question that matters is whether (permanent) changes in M *cause* (offsetting permanent) changes in V. (We both recognise, of course, that expected increases in the growth rate of M will cause increases in V.)

[Too Much Fed]

Mike Sproul said: "The fed might refuse to allow reflux of either its paper money or its deposit money. Either way, it amounts to a default by the fed. This is tantamount to a loss of backing, and can be inflationary. But there are qualifications: Maybe the fed will resume reflux in 1 year, or maybe it will pay interest on its paper money or on its reserves until reflux resumes. In that case the effect of plugging a few reflux channels can be negligible"

The way the system is set up now paper money (currency) can always reflux. The fed's deposit money (central bank reserves) may or may not reflux. Those qualifications are important.

Lorenzo said: "Change the medium of account to Central Bank base money. Why is its value not set by supply and demand for it?"

Lorenzo and Mike S., let's change MOA to currency plus demand deposits. Now the fed increases central bank reserves and demand deposits. Next, people repay bank loans with demand deposits so currency plus demand deposits stays the same. Not much happens?

How does the backing theory work if gov't bonds are backing currency and central bank reserves?

[Mike Sproul]

Too much fed:

"paper money (currency) can always reflux. The fed's deposit money (central bank reserves) may or may not reflux."

Think of a leaky vacuum cleaner hose. You plug one hole and the air can no longer reflux through that hole, so it will reflux through another hole instead. On Monday people might bring paper dollars to the Fed as a loan repayment and the Fed might accept them. On Tuesday the Fed might stop accepting paper dollars for a day or two, but the fed might make an open market sale of its bonds in exchange for its own deposit dollars. Reflux can happen at the instigation of the fed, or at the instigation of the public. It's still reflux either way.

"let's change MOA to currency plus demand deposits. Now the fed increases central bank reserves and demand deposits. Next, people repay bank loans with demand deposits so currency plus demand deposits stays the same. Not much happens?"

It's true that not much happens, but I would never lump currency and demand deposits together like that. Currency was issued by the fed and is the fed's liability. Demand deposits are Wells Fargo's liability and were issued by Wells Fargo. They can't be lumped together any more that GM bonds and Ford bonds can be lumped together.

[Min]

Err, in the US, it is the Treasury that issues currency.

[Mike Sproul]

Min:

That requires more explanation. For example, why are paper dollars called "Federal Reserve Notes", and why do those FRN's show up on the liability side of the Fed's balance sheet? Of course this would be a moot point if we combined the balance sheets of the Fed and Treasury.

[Max]

Mike, "If cover is truly .9 oz per dollar, then each dollar must be worth .9 oz."

Cover is another word for reserves. I don't think you mean to say that the value depends on the reserve ratio.

[Min]

@ Mike Sproul

See http://www.law.cornell.edu/uscode/text/31/5115 and http://www.law.cornell.edu/uscode/text/31/5111

:)

[Too Much Fed]

Mike S. said: "Currency was issued by the fed and is the fed's liability."

I'm going to let that one stand for this conversation.

"Demand deposits are Wells Fargo's liability and were issued by Wells Fargo."

OK.

"They can't be lumped together any more that GM bonds and Ford bonds can be lumped together."

GM bonds and Ford bonds are not fixed convertible to each other or to currency and demand deposits. Currency and demand deposits are fixed convertible.

Let's say currency is MOA. Now have a fixed conversion rate between currency and demand deposits. Is there a dual MOA now?

[Too Much Fed]

"On Monday people might bring paper dollars to the Fed as a loan repayment and the Fed might accept them. On Tuesday the Fed might stop accepting paper dollars for a day or two, but the fed might make an open market sale of its bonds in exchange for its own deposit dollars. Reflux can happen at the instigation of the fed, or at the instigation of the public. It's still reflux either way."

I don't think people bring paper dollars (currency) to the fed as a loan repayment. I'd say this usually happens: Some entity does not want currency anymore. It takes the currency to a commercial bank to get demand deposits. The commercial bank can then take the currency to the fed for central bank reserves. The reflux of currency can always happen. The Tuesday part above does not happen. It would violate the elastic currency part of the Federal Reserve Act (at least that is how it was explained to me).

The commercial banks may want to reflux the central bank reserves back to the fed. The fed may or may not allow that reflux part.

[Too Much Fed]

Min said: "Err, in the US, it is the Treasury that issues currency."

Is currency ever a liability of the Treasury/Bureau of Engraving?

[Mike Sproul]

Min:

That link you provided discusses Treasury-issued currency, but the paper money that we use every day is issued by the Federal Reserve.

[Mike Sproul]

Too much Fed:
"GM bonds and Ford bonds are not fixed convertible to each other or to currency and demand deposits. Currency and demand deposits are fixed convertible."

Wells Fargo chooses to make its checking account dollars convertible into Fed dollars, but the Fed does not make its dollars convertible into WF dollars. Also, WF dollars are WFs liability, and fed dollars are the fed's liability. But the Fed's paper dollars and the fed's deposit dollars are BOTH liabilities of the fed.

If a private bank has unwanted central bank reserves, it can swap them for paper dollars, which you say can reflux to the fed. It's the leaky vacuum hose again.

[Mike Sproul]

Max:

My mistake. I was using "cover" to mean "assets"

[Lorenzo from Oz]

NIck: an unexpected way of presenting it, but helpful, thanks.

Rene: Yes, it was understood that was Peel's intent. It is still very unusual.

Mike Sproul: This is beginning to look like a case of shifting goal posts. If the cover ratio determines the value, why did the value of currencies not move around according to the cover ratio? As it happens, the exchange rate with gold mattered directly, not as adjusted for cover ratio. If the cost of remaining on the gold standard (typically in outflow of gold) got too difficult, countries simply left the gold standard. The credibility of the promise to exchange is what mattered, not the cover ratio per se. The value of the franc did not change as the BoF stockpiled gold (1928+) and the cover ratio got higher and higher. Indeed, gold continued to flow in precisely because the franc remained "undervalued", just as gold flowed out of the UK because the pound was "overvalued".

On the peso case, that presumes every monetised transaction is up for use by any money. Clearly, not how it works. Network effects are very strong, since money needs expected transaction chains to retain value. That is a large part of why CSA$ spectacularly hyper-inflated: the expected duration of the transaction chains kept getting shorter and shorter ...

[Min]

Too Much Fed: "Is currency ever a liability of the Treasury?"

The US Treasury accepts US dollars in payment of taxes. :)

[Min]

Mike Sproul: "the paper money that we use every day is issued by the Federal Reserve."

I showed you mine. You show me yours. What act of Congress authorizes the Fed to issue currency?

Besides, do you think that the bills and coins issued by the Treasury do not circulate in the economy?

[Nick Rowe]

Lorenzo: here's a second way to think about Mike's perspective: Mike wants to liken money to shares in (say) a mutual fund. But suppose the directors of a closed end mutual fund used their discretion to pay varying amounts to charity (or the government) from the income earned from the assets of the fund, rather than returning it all to the mutual fund shareholders. Those shares would not trade at Net Asset Value.

[Mike Sproul]

Lorenzo:

I'm still paying for saying "cover" when I meant "assets". It's assets that determine value, not cover. (At the risk of misusing words again, cover=reserves)

So if the fed had 100 oz worth of assets as backing for $100, then $1=1 oz. Complications arise quickly. If the fed's assets are worth 140 oz, but the fed maintains convertibility at $1=1 oz, then the dollar will still be worth 1 oz, since the fed is refusing to use its extra 40 oz to back its dollars, and so the 40 oz might as well be on the ocean bottom.

On the other hand, if assets fell to 90 oz, and the fed maintained convertibility at $1=1 oz, then there would be a run on the bank, and you get a situation like you described in the UK.

"money needs expected transaction chains to retain value."

Not if the money is backed. In the case above, $1=1 oz whether or not the dollar is used as money, since you can always get 1 oz worth of assets from the issuer of the dollar.

[Jacques René Giguère]

Nick:"It is still very unusual." And that's the point. Since 1884, the British discussed and understood the limits of the system, just where you can go, when to intervene before it is too late so you don't need to break the rules and so on.
Something the BCE didn't do and the Franco-Germans don't even begin to understand.
In the end, Texans are right: better be judged by twelve than carried by six.

[Nick Rowe]

Jacques Rene: you are responding to Lorenzo, not to me! (Names are below the comments.)

[Jacques René Giguère]

Nick: my old eyes again . One day everyone will be as old as I am and this blog will be a paradise of cross-talk...

[Lorenzo from Oz]

Mike Sproul: if assets fell to 90 oz, and the fed maintained convertibility at $1=1 oz, then there would be a run on the bank, and you get a situation like you described in the UK
But currencies did not have runs on them merely because the cover ratio expanded. What people cared about was whether the commitment to exchange notes for gold/silver/specie was credible. Currency exchange rates would tend to hover within "gold points" (the rates at which it was worthwhile transporting gold rather than notes). The problem in the UK in 1926-31 was that the rate of exchange of UKP for gold was set too high (i.e. pounds were overvalued in terms of gold) so gold would buy more outside the UK. (Conversely, the franc was undervalued in terms of gold, so gold would buy more in France, and so flowed in, driving up the franc's cover ratio, which never hit 100%.) A problem that worsened as the BoF, aided and abetted by the Fed, drove demand for gold up. When it appeared that the BoE would run of out gold to make the exchange rate credible, then the UK went off the gold standard.

Not if the money is backed. Were cowrie shells "backed"? Because they were by far the most widespread form of money, across time and space. But history is full of (sometime unexpected) local monies, because they were what had high transaction utility in local transaction chains.

Spanish "pieces of eight" were used as money in Chinese maritime provinces extensively from 1565 onwards, typically exchanging at higher value than weighted silver because the "brand" was so reliable.

So perhaps backing theory only applies to paper money. But, as I am intimating, a gold/silver/specie standard worked rather more like fractional reserve banking: as long as confidence was fine, notes could be issued well in excess of the gold/silver/specie "backing".

(I keep saying gold/silver/specie because the fetishisation of gold annoys me. Silver was a much more important monetary metal than gold and there is a good argument that the classical gold standard was less stable than the some-on-gold, some-on-silver, some-on-specie structure that preceded it.)

Nick: nice try, but I am not a finance brain, so that does not help.

Rene: Matthew Morys paper on Monetary Policy Under the Classical Gold Standard suggests that the Reichsbank may have been more central in running the 1873-1914 system that folk have previously thought.
http://econ.as.nyu.edu/docs/IO/18864/Morys_20110204.pdf

[Jacques René Giguère]

Lorenzo: as I say to my admiring students, If I seem to know everything, it's because I learn everything. So mucho thanks for the link.
Which begs the question: what made them lose their proficiency?

[Mike Sproul]

Lorenzo:

1. I want to stop using the word "cover". The backing theory says it's assets that matter, not cover. If assets are enough to buy back every unit of money issued at par, then people will believe that a convertibility promise is credible.

2. Cowries are a commodity. The backing theory doesn't apply. The backing theory applies to moneys that are the liability of the issuer.

3. How big was the premium on pieces of eight? Any significant premium would be arbitraged away.

4. I agree about silver vs gold.

[Lorenzo from Oz]

Jacques Rene: Excellent question. By the time of the hyperinflation, the Reichsbank was in la-la land. But the "real bills" doctrine possibly makes more sense when your other experience is gold/silver/specie standard. Indeed, as long as the offer to exchange is credible, then issuing notes for "real bills" is fine.

More recently, it is clear that the Euro-elites hated floating exchange rates. (Down here in Oz, we love them. But we don't aspire to signing up to some ever-closer union.)

Marc Flandreau and others published a paper in 1998 which looked at the gold standard and which warned that it would not work if the ECB was too restrictive.
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=108211

BTW, have you seen this paper? It has amusement value (in a black humour sort of way). Particularly given the timing of its release.
http://ec.europa.eu/economy_finance/publications/publication16345_en.pdf
A nice summary by a political scientist is here:
http://www.craigwilly.info/2012/07/28/we-told-you-so-how-u-s-economists-predicted-the-euro-crisis/

Mike Sproul: Well, yes, the gold/silver/specie standard was "backed". But in a straightforward exchange-for-metal way in the "core" countries. (The "periphery" countries were a bit different.) If the cover ratio was greater than 1:1, then that is a problem for backing theory, surely? Unless it is to be restricted to money's not up for explicit exchange.

On the Spanish dollar premium in China, it varied, but sometimes got quite high. A paper on the matter reports the following:
"However already in 1797 Morse reports that they were at par with sycee [weighted silver]. From 1811 the same source indicates that the coins enjoyed a premium over sycee despite their lower content of fine silver, and in 1829 the premium was already around 6 to 8 %. ... The premium only continued rising in the 1830s to 14 and 15%, and in the 1850s it was 20-30%, as Eduard Kann reports for 1852. The rate overshot at the time of the Taiping when it went up as high as 50%, yet in 1863 the coin still enjoyed a premium of 15% upcountry."

The paper by Alejandra Irigoin is available here: http://eprints.lse.ac.uk/49082/

The author makes the point that it was demand for a currency standard (i.e. a medium of account) which drove the premium, remembering that the Chinese state issued no silver coins, and took taxes in weighted silver. It was, of course, possible to separate the unit of account and the medium of exchange (it was done a lot when there was a lack of reliable coin) but it is not very convenient.

Money is a very particular type of asset--its transaction utility is not only tied to being widely exchangeable but also use as a unit of account. It is both an asset and a network good. Makes it odd.

[Too Much Fed]

Mike S. said: "Wells Fargo chooses to make its checking account dollars convertible into Fed dollars, but the Fed does not make its dollars convertible into WF dollars."

Let's say a lot of demand deposits are created from banks. Next, everyone redeems their demand deposits for currency. As long as the banks are solvent, the fed will allow this increase in currency. The fed does not run out of currency. The fed does not disallow this exchange. If somehow a lot of currency is created, an increase in demand deposits could happen if people redeemed their currency for demand deposits. That is what is important.

"Also, WF dollars are WFs liability,"

OK.

"But the Fed's paper dollars and the fed's deposit dollars are BOTH liabilities of the fed."

I'm going to skip the paper dollars (currency) part for now. The fed's deposit dollars (what I call central bank reserves) are liabilities of the fed.

"If a private bank has unwanted central bank reserves, it can swap them for paper dollars, which you say can reflux to the fed. It's the leaky vacuum hose again."

For a bank, central bank reserves and currency can be swapped at the fed. The banks can't force the fed to give up some assets in exchange for central bank reserves and/or currency if the banks have too many central bank reserves and/or currency.

[Too Much Fed]

Min's post said: "Too Much Fed: "Is currency ever a liability of the Treasury?"

The US Treasury accepts US dollars in payment of taxes. :)"

I'm not sure that is answering my question. The US Treasury accepts demand deposits for payment of taxes too.

[Jacques René Giguère]

Lorenzo: once again,thanks. As I sometimes acknowledge, I am an humble IO guy. I come on monetary thread to get a lot of knowledge and, I hope, a bit of wisdom. I gained a lot in the last few hours. As Walpole said "Life is a tragedy for those who feel and a comedy for those who think"...
The French seems to have been AWOL on economic thinking since the days of Say, Bastiat and Cournot. Even in the '60, De Gaulle was making long speeches about gold, a great military tactician and stategist, a brilliant politician caught in a "ne supra crepidam" hell. As for the Germans, you just gave us some insights.
This being said. Mundell was in favor of the euro. Seemingly because he was also a hard-money type and a social conservative who enjoyed the idea that a neo-gold standard would let the working class suffer the downward adjustments while the anti-inflation clauses of the pact would take care of any improving wage situation.
Optimism? Well, polito-bureaucrats had seen French-German enmity disappear, the Cold war blocs vanished and, almost within their living memory,seen France complete its economic unity and what where at the time seemingly impossible feats, the creation of a somewhat workable Italy and a (maybe too much) well-functionning Germany out of thin air. Could we fault them for optimism? What will be history's judgment a century hence, remembering that Germany very nearly broke up in 1918-19?

[Lorenzo from Oz]

Jacques Rene: Well, yes, there certainly was a lot of optimism. And perhaps sheer political bloodymindedness will make the Euro work. BTW, the comment on the paper by Flandreau et al should have said "they warned that the Euro would not work if the ECB was too restrictive".