Sunday, September 28, 2014

The law of reflux

One of the coining press rooms in the Tower of London, c.1809 [link]

[This is a guest post by Mike Sproul.]

The law of reflux thus assures the impossibility of inflation produced by overexpansion of bank credit. (Blaug, 1978, p. 202.).

It is the reflux that is the great regulating principle of the internal currency; and it was by the preservation of the reflux, throughout all the perils and temptations of the period of the restriction, that the monetary system of these kingdoms was saved from the utter wreck and degradation which overwhelmed every paper-issuing state on the Continent… (Fullarton, 1845, p. 68.)

If you want to understand the law of reflux (and you should), then think of silver spoons. The silversmith shown in figure 1 can stamp 1 oz. of silver into a spoon. If the world needs more spoons, then silversmiths will find it profitable to stamp silver into spoons. If the world has too many spoons, then people will find it profitable to melt silver spoons. Unwanted spoons will “reflux” back to bullion. In this way, the law of reflux assures that the world always has the right amount of silver spoons. We could hardly ask for a simpler illustration of the Invisible Hand at work. But it costs something to stamp silver and to melt it, so the price of spoons will range within a certain band. It might happen, for example, that silversmiths only find it profitable to produce spoons once their price rises above 1.03 oz., while people only find it profitable to melt spoons once their price falls below 0.98 oz. If the costs of minting and melting were zero, then a spoon would always be worth 1 oz.

This is a point worth emphasizing: The value of a spoon is equal to its silver content. An increase in the demand for spoons would not raise the price of spoons much above 1 oz, since new spoons would be produced as soon as the price rose above 1 oz. A drop in the demand for spoons would not push the price much below 1 oz, since spoons would be melted when the price fell below 1 oz. Likewise, if the quantity of spoons supplied became too large or too small, market forces would restore the quantity of spoons to the right level, while keeping the price at or near 1 oz.

In figure 2, the silversmith starts being called a mint, and instead of stamping silver into spoons, the mint stamps silver into 1 oz. coins. The law of reflux works the same for coins as for spoons, always assuring that the world has the right amount of coins, and that the value of each coin always stays at 1 oz., or at least within a narrow band around 1 oz.

The usual thought experiments of monetary theory don't work in this situation. For example, economists often imagine that if the money supply were to increase by 10%, then the value of money would fall by about 10%. But the law of reflux won't allow this to happen. Mints would only issue 10% more coins if the public wanted coins badly enough to part with an equal amount of their silver bullion. And even if mints went against their nature and issued more coins than the public wanted, those coins would be melted or stored, and the value of a coin would not deviate very far from its silver content of 1 oz.

In figure 3, the mint re-invents itself again, this time as a bank. Rather than stamping customers' silver into coins, the bank stores the silver in a vault, and issues a paper receipt called a bank note. (Checkable deposits would also work.) This system has several advantages over coins. (1) It saves the cost of minting and melting. (2) It avoids wear of the coins. (3) Bank notes are harder to counterfeit, easier to carry, and easier to recognize than coins.

The law of reflux still works the same for bank notes as it did for coins. If the economy is booming and people need more bank notes, then people will deposit silver into their banks and the banks will issue new bank notes. If the economy slows and people need fewer bank notes, then people will return their unwanted notes to the banks and withdraw their silver.

Once again the thought experiment of imagining a 10% increase in the quantity of bank notes is pointless. Banks would only issue 10% more notes if the public wanted those notes badly enough to bring in 10% more silver. And even if we imagine that the banks took the initiative, printing 10% more notes and using those notes to buy 10% more silver, every bank note is still backed by 1 oz. and redeemable into 1 oz of silver at the bank, so every bank note remains worth 1 oz.

In figure 4, the bank makes one more important change. Rather than requiring customers to bring in 1 oz of actual silver to get a bank note, the bank also accepts an equal or greater value of bonds, bills, real estate deeds, or anything else that can fit in the vault. This system has several advantages: (1) Handling bonds, etc. is easier and safer than handling silver. (2) The silver formerly deposited can be put to productive use. (3) The quantity of bank notes is no longer constrained by the amount of silver available. (4) The bank earns interest on its bonds, bills, etc.

The law of reflux still operates as before, except that when people want more notes, they can bring in either silver or bonds, and when people have excess notes, the notes can be returned to the bank for either silver or bonds. As before, it makes no sense to ask questions like “What if the bank issues 10% more bank notes?” And as before, so long as every bank note is backed by, and convertible into, 1 oz. worth of assets, every bank note will be worth 1 oz. Just as reflux assures that the value of a spoon is equal to its silver content, reflux also assures that the value of a bank note is equal to the value of the assets backing it.

The Channels of Reflux

Here is a list of some of the many channels through which bank notes might reflux to the issuing bank:
1. The silver channel: Unwanted notes are returned to the bank for 1 oz. of silver. Alternatively, the bank sells its silver for its own notes, which are retired.
2. The bond channel: The bank sells its bonds in exchange for its notes, which are retired.
3. The loan channel: The bank's borrowers repay loans with the bank's own notes.
4. The real estate channel: The bank sells its real estate holdings for its own notes.
5. The rental channel: The bank owns rental properties, and tenants pay their rent in the bank's notes.
6. The furniture channel: The bank sells its used furniture for its own notes.

As long as enough reflux channels are open, it does not matter if a few channels are closed. Customers would not care if the furniture channel was closed, as long as major channels, like the bond channel, stayed open. The bank could take a more drastic step and close the silver channel, or could delay silver payments by 20 years, and as long as enough other channels stayed open, the law of reflux could operate as always, except that notes might be redeemed for 1 oz. worth of bonds, rather than 1 oz. of actual silver. The bank could even un-peg its notes from silver. Rather than redeeming a refluxing dollar note for 1 oz. worth of bonds, it could redeem dollar notes for 1 dollar's worth of bonds. As long as the bank's assets are worth so many oz., it doesn't matter if those assets are denominated in oz. or in dollars.

Once metallic convertibility is suspended, it would be an understandable mistake if people forgot all about the other channels of reflux, and started to think that bank notes were no longer backed by, or convertible into, anything at all. Unfortunately, this mistake has made it into the textbooks:
You cannot convert a Federal Reserve Note into gold, silver, or anything else. The truth is that a Federal Reserve Note has no inherent value other than its value as money, as a medium of exchange. (Tresch, 1994, p. 996.)
There you have it. The closing of just one channel of reflux (the metallic channel), has fooled economists into wrongly rejecting the idea that modern bank notes like the US paper dollar are backed and convertible. Once economists reject this simple and obvious explanation for why modern paper money has value, they are forced to resort to the more exotic explanations offered by textbook monetary theories, which are anything but simple and obvious.

Blaug, Mark, Economic Theory In Retrospect, 3/e. Cambridge: Cambridge University Press, 1978
Fullarton, John, Regulation of Currencies of the Bank of England (second edition), 1845. Reprinted by Augustus M. Kelley, New York: 1969.
Tresch, Richard, Principles of Economics, St. Paul, Minnesota: West, 1994


  1. The medium of exchange is not like other assets. It is not like silver spoons. We borrow money, not because we want to hold more money, but because we want to spend it. The person who accepts it from us does not want to hold it either. He in turn wants to spend it. And so on.

    1. True; but nevertheless, aggregate cash holdings rise and fall with the state of the economy. Mints only issue more coins to people who bring in silver, and banks only issue new notes to people who bring in assets of adequate value. As long as 10% more money is accompanied by 10% more backing, money holds its value.

      Whether or not you buy the whole idea of the law of reflux, you have to admit that economists make a big mistake when they claim that the suspension of metallic convertibility is tantamount to a complete loss of backing.

    2. "The medium of exchange is not like other assets. "

      It sounds like you're denying Mike's point about coins. Given that a coin (the "medium of exchange") is is more liquid than a spoon, this doesn't change what is basically a relationship driven by arbitrage. As long as something can be costlessly turned into something else (either physically or via redemption) then the price of spoons, silver bullion, and coins cannot diverge in any meaningful way. Introducing costs drives a wedge between the two prices, like minting fees, spoon pressing fees, costs of promoting liquidity, etc.

  2. Thanks Mike, excellent post. I would really like to see economists to engage more with all your work. IMO it should garner more interest.

    1. Jussi, Mike used to debate a lot with die hard gold standard fans over at (I was one of them, a Misesesian for sure having read thousand page tomes by Mises and Rothbard) and I became interested in his ideas in 2007 when he had this great lively debate in the comments section of this article. Unfortunately it looks like has cleaned their site and scrubbed old material. Pity because that was one of the best debates I've ever seen. After that, I read all of his papers and threw out the idea of fiat money and the need for a gold standard.

  3. This seems simple but deep.

    I suppose a corollary of this is that central banks tend to have balance sheets with positive net asset value.

    Does the fact that central banks tend to hold bonds that reflect a cumulative government deficit have any bearing on the backing view?

  4. JKH:

    I don't see your point about positive net asset value. A central bank that issued $100 of bank notes in exchange for bonds worth $100 would have zero net asset value.

    About cumulative government deficits: In the simplest case, the central bank that issues $100 might issue them for $100 worth of its own government's bonds, or some other government's bonds, or corporate bonds, real estate, or lottery tickets. All that matters is that the central bank gets assets of adequate value for the money that it has issued. Of course, if the government deficit explodes later on, then the government bonds owned by the central bank will lose value, and the backing view would predict inflation.

    I have a feeling you might have something else in mind though...

    1. I could have said "non-negative net asset value" I guess. That is a sort of evidence that something of (adequate) value has been received in exchange for money issued. The fact that central banks tend to be capitalized above the point of zero net asset value is beside that point, but factual and reinforcing.

      When the standard central bank balance sheet is consolidated with the standard government treasury position, the standard result is a net liability position - negative net asset value from that perspective. So the backing for money disappears on that basis. That said, maybe taxation power becomes backing in that sense. This all may "just" be accounting, but I think it says something - maybe something about backing that is congruent with the accounting.

    2. JKH:
      Re consolidated balance sheets:
      1) Even if net worth of the Fed and Treasury combined is negative, Federal Reserve Notes are a "First and paramount" liability of the fed. If FRN's are first in line for the combined assets of Fed/Treasury, then they can be fully covered even if combined net worth is negative.
      2) Questionable if combined net worth is negative, once you add up all the land, taxes receivable, etc.
      3) Taxation power has often served as backing. It's like the "rental channel" I mentioned in the article. In the case I mentioned, "rents receivable" is an asset to the bank, and serves as both backing and as a reflux channel. Change "rents receivable" to "taxes receivable", and it's clear that if the fed/treasury are combined, then 'taxes receivable' serve as backing and reflux channel for FRN's. Even if they aren't combined, taxes still back the fed's bonds, and the fed's bonds back FRN's.
      4) I've learned to trust accountants way more than I trust economists. The backing theory is fully consistent with accounting theory, while the quantity theory breaks every accounting rule in the book, starting with quantity theorists' denial that the fed's liabilities are really liabilities.

    3. So if central banks have positive net asset value and combined with goverments they also have positive net asset value then where does inflation come from? If a bank has 120 assets, 100 liabilities and 20 in owners equity then a loss of 10 would not effect the value of the liabilities. Neither would a gain of 10. Somehow central banks around the world manages to get their liabilities valued lower every year.

    4. Dan:

      Inflation happens either because the central bank has less assets relative to liabilities, or because the central bank decides to devalue its liabilities. Sometimes central banks lose 10% of their assets, or their liabilities grow by 10%, and so they get something like 10% inflation. Other times, central banks just announce that each dollar they issued will henceforth be redeemed for 10% less gold than before. Other times, the central bank announces that it will pursue a 10% inflation target, effectively defaulting on 10% of their liabilities each year, even though their assets might have been enough to keep their dollars at par.

    5. So assuming a central bank that manage to keep a good margin of positive net asset value then all inflation comes from devaluing the liabilites of the central bank, not from changes in total liabilites or assets, correct? Because assuming positive net asset value then changes in backing would only affect owners equity. And how do a modern central bank devalue its liabilities? Just by announcing it? Wouldn't they have to take active steps to lower the interest rate? Which they do through open market operations by buying bonds. So how can that be considered defaulting on some of the debt? Just trying to understand.

    6. Dan:

      Here's an easy case, and I'll see if I can get to a more detailed answer later this week.

      In 1797, the Bank of England suspended gold convertibility, and they resumed it in 1821. Whenever the BOE announced a rate and date for convertibility resumption, the pound would move in anticipation of that expected future rate. It had nothing to do with the BOE's rate on loans or its open market operations. But let's say the pound just fell 10%. Then people will need 10% more pounds to conduct their usual business, so they will either come to the bank asking for loans or else offering to sell bonds/bills to the bank. The BOE would be inclined to issue new money in these times, since they would otherwise face the squawks of businesses suffering from tight money. A quantity theorist would notice that the 10% fall of the pound was accompanied by a 10% rise in the quantity of pounds, and would think he had found evidence supporting his view. Thomas Tooke pointed out way back in 1840 that inflation always preceded rises in the money supply, and correctly concluded that the real bills/backing view was correct. That's why his name has been dragged through the dirt ever since.

    7. "...then all inflation comes from devaluing the liabilites of the central bank, not from changes in total liabilites or assets, correct?"

      Too some extent, inflation is a natural phenomenon.

      Consider a highly liquid asset that provides a competitive return. By virtue of its excellent marketability, a large part of this asset's return comes in the form of a liquidity return. Investors will therefore be satisfied with only small year-over-year capital gains. As the good becomes even more liquid, investors may even be willing to accept a slightly negative year-over-year fall in that asset's price in order to enjoy its incredible liquidity return. We could argue that central bank notes and reserves are like this. They are so liquid that people accept a -2% year-over-year capital loss on them.

      A central bank can tweak this expected capital loss rate (ie inflation) via open market operations. And because it has plenty of backing, there's no reason for the value of those liabilities to fall, say, by 10% overnight.

    8. Thanks Mike, I can understand the BOE example, are you also saying that in todays modern systems the act of the central bank announcing a change in inflation policy will make the market adjust their valuation accordingly without any further action needed? I 'm thinking about Japan and how they've struggled to reach their 2% target even though this policy has been well communicated. On the other hand I feel it would be rational for the market to make some move after an announcement of that kind.

      JP, I remember reading this now:

      So I'm still not sure how to think about liquidity regarding inflation. I understand people are willing to take a small loss for a highly liquid asset but how is this loss created is the question. Increased liquidity has a diminishing effect the more it is done so it shouldn't work as a means to continueously create inflation by itself. So maybe the CB can get the market moving (valuing money lower) by adding liquitity but the price of money is sustained at that lower level because the CB will not let any money reflux back into its bonds from that new lower price level of money. Effectivly this would mean defaulting on part of it's liabilities like Mike said. Then from that lower price level demand for money once again increases so that extra liquidity can continue to have an effect when it's added. Something like that?

      Thanks for interesting answers.

    9. Dan:

      In 1797 the BOE suspended gold convertibility for 24 years, but maintained bond/loan/furniture convertibility the whole time. The fed has suspended gold convertibility for 80 years and counting, but has kept the other channels open. So there's not much difference between the BOE system and modern systems. Back then the BOE would announce a future gold peg. Today the Fed announces a peg to the CPI, but it works the same. I should emphasize that the announcement must be CREDIBLE.

    10. "...but how is this loss created is the question."

      Imagine that the economy-wide riskless rate of return is 2% so that adjusted for risk and liquidity, all assets will yield 2%. At the outset, assume that Fed reserves provide investors with a liquidity return of 2.5%. This means that the market will expect the price of reserves to fall by 0.5% each year (2.5 + -0.5 = 2%). To execute policy, the Fed simply modifies this expected loss by making slight increases to the original 2.5% liquidity return. For example, if it tightens the quantity of reserves, than the marginal reserve now provides more liquidity, say 2.6% rather than 2.5%. Investors will heretofore accept an expected 0.6% loss (2.6% + -0.6 = 2%) on reserves, more than the initial 0.5% loss. This tighening will create some deflation since one way that market expectations of a larger loss (0.1%= 0.6% - 0.5%) are established is by a one time jump in the value of reserves (or a one time fall in the price level) relative to its expected resting point one year hence.

      I'm not sure how much Mike would agree with that.

    11. Mike, if I for my own sake describe things like this, would you agree?:

      The Fed is prepared to buy or sell bonds (or other things of value) to make the value of the currency fall in line with the inflation peg and the market will adjust to this fact before any action by the Fed is taken as long as they know about the peg and it is credible. And because of this there is no need for the Fed to do OMO's to execute inflation policy but rather to avoid tight money conditions after the market already has adjusted to a higher price level.

      To me it seems that there is convertibility but it is entirely controlled by the Fed. The market can't come to the Fed requesting som backing for dollars. They just have to anticipate what the Fed does and value the currency accordingly. If the Fed creates more money than the market expects it might not lead to inflation because the market then expects the Fed to hold it's peg and take some of the table again or slow down in the future. I'm not sure how the loan reflux channel works though.

      JP, thanks for the explanation. You then think that OMO's are in fact needed to get the market to value the currency in line with the stated inflation goal, I must assume?

    12. JP:

      I agree. The most direct answer is that if the Fed wants 2% inflation, even though its assets could support 0% or less, then the Fed issues more money, either on loan or through buying bonds. That extra cash wants to reflux to the Fed, but the Fed, in keeping with its 2% target, closes the reflux channels just enough that the money supply stays at its artificially high level. That could reduce the value of the dollar in a couple of ways. 1) You and I often talk about the dollar having a 5% or so premium over its backing value. That excess money could drive the money down in value by 2% by reducing the monetary premium. (2) People will expect that when the Fed finally does redeem its dollars for a physical basket of goods, the basket will be 2% smaller, in keeping with the Fed's 2% inflation target.

    13. Dan:
      Correct. It doesn't matter if convertibility happens at the Fed's initiative or at the public's initiative. The Fed might announce to the public that it always stands ready to redeem $1 for 1 oz, and let the public decide to redeem or not, or the Fed might have a policy that whenever the dollar falls below 1 oz, the fed sells silver and buys $, while whenever the dollar rises above 1 oz, the fed will sell $ and buy silver. Regardless of who takes the initiative, $1=1 oz.

      Loan reflux is no different. It makes no difference if people decide to repay loans or if the fed calls in the loans. The lent money refluxes to the fed either way.

    14. Mike, ok great. And if the Fed credibly raises it's inflation target from 0 to 2% but for some reason don't issue any more money, will there be inflation and a tight money situation or will there be no inflation?

    15. Dan:

      Correct. Inflation and tight money. Here's a related quote from Hayek describing the BOE restriction period of 1797-1821:

      "It is not easy to reconcile these complaints about the continued scarcity of money during this period with the no less insistent complaints about high prices, and with the continued unfavorable course of the exchanges.” (Hayek, 1933, p. 40)

  5. If a bank with 100 % silver reserves wants to issues new silver notes it has to face the market prices of silver, which it cannot really influence. If fractional reserve bank creates new money and lends it out it actively changes the price of all the money by increasing its supply. It’s like making new spoons without having to get the silver from the market first. You’re not just changing the form of the silver from bullion into spoons, you actively increase the overall supply of both. In that case the law of reflux cannot hold, because the constraining forces of physical scarcity on your business are just not there. At zero marginal cost of a new spoon you can always underbid the next seller of spoons and flood the market with your product. Melting your spoons back into bullion following the increase in its price relative to more abundant spoons will decrease the bullion prices accordingly. But this reflux of spoons back into bullion cannot stop you from producing even more spoons as you are not constrained by the physical market anymore. Today’s money producer (for example bank with fractional reserves) can by its actions change the price of all money in the economy while the bank with 100 % reserves was just changing relative prices of various forms of money (spoon/silver, silver bullion /silver coin, bank note /silver bullion….)led by arbitrage profits in between them.

  6. Juraj:
    "If fractional reserve bank creates new money and lends it out it actively changes the price of all the money by increasing its supply."

    This assumes its own conclusion. According to the backing theory, new money issued on fractional reserves will NOT change the price of money, since the new money is adequately backed by the new assets that the bank buys with that money.

    You accept that new notes issued in exchange for silver would not reduce the value of the notes. Take another step and suppose that additional notes are issued, not for 1 more oz of actual silver, but for an equivalent value of gold (i.e., an amount of gold worth 1 oz of silver). It's pretty clear that this will not reduce the value of the notes either. Now issue new notes for an amount of wheat worth 1 oz, or for bonds worth 1 oz. The bank still has enough assets to buy back all its notes for 1 oz of silver or its equivalent in other goods or bonds, so the money still holds its value.

    1. Mike, I am not sure I understand the reason, why the relative price of monetary unit should depend on the value of “backing” and not on the relative quantity of monetary units to other goods in the economy. There is no fixed exchange rate between notes and “backing” assets anymore. I see inflation also in your suggested next steps. If silver dollar is money (everything is priced in silver dollars), than leaving the exclusive silver backing behind (for example 1 silver dollar = 1 oz. of silver) and issuing additional silver dollars backed by gold (bonds, wheat, ....) is inflationary, because now there are more monetary units in circulation and the same number of apples, cars, and employees in the economy. The price of silver dollar in other goods will fall, because original prices based on the available silver as the monetary metal have to increase to reflect the inclusion of other assets into the monetary sphere as well. It’s a process similar to finding new silver deposits in the ground. Yes, the bank can take the new silver dollars out of the system, If it wishes to do so (assuming no losses on backing assets). but till it does, there will be an increase in the money supply.

    2. Juraj:
      1) “Mike, I am not sure I understand the reason, why the relative price of monetary unit should depend on the value of “backing” and not on the relative quantity of monetary units to other goods in the economy.”

      Everyone agrees that the value of stocks and bonds depends on their backing, and not on their quantities relative to other goods. For example, there might be 1 million shares of GM stock in existence, and the value of GM’s assets might be $60 million. The implied share price is $60. If GM then issued another 1 million shares and sold them in exchange for another $60 million worth of assets, then there would be 2 million shares backed by $120 million of assets, so shares are still worth $60 each, even though the quantity of GM shares has risen relative to goods in the economy.

      But back to money. Initially $100 are backed by 100 oz of actual silver, so $1=1 oz. Then the bank issues another $20 in exchange for gold that is worth 20 oz of silver. Suppose that this makes the street value of $1=0.8 oz of silver. An arbitrager could then buy all $120 of circulating dollars for about 100 oz of silver, and he owns the bank, which contains 100 oz of silver plus 20 oz worth of gold. Arbitrage assures that $1 must be worth 1 oz. There can be no “dollar inflation” in the sense of the dollar falling relative to silver.

      The confusing thing about this, and the thing that I was stuck on for a few years, is that even though there can be no dollar inflation, there can be “silver inflation”. As paper money displaces silver money in peoples’ pockets, the demand for silver falls and silver will buy less bread than before, even though it’s still true that $1=1 oz of silver. There has been silver inflation, but no dollar inflation.

      2) “Yes, the bank can take the new silver dollars out of the system, If it wishes to do so (assuming no losses on backing assets). but till it does, there will be an increase in the money supply.”

      If dollars are free to reflux to the bank, then any fall of the dollar relative to silver will prompt a reflux of silver to the bank, and any dollar inflation will be cut off before it starts. If dollars are not free to reflux, then that is tantamount to a default by the bank, and the dollar will lose value because it has lost backing.

      3) “There is no fixed exchange rate between notes and “backing” assets anymore.”

      What matters is backing, not exchange rates. If the bank has 100 oz backing $100, then $1=1 oz. If the bank declares an exchange rate of $1=1.2 oz, then that declaration will prompt a run on the bank. The first $80 through the bank’s door will be redeemed for 1.2 oz each, and the last $20 will get nothing.

    3. Mike: If GMs shares were money (goods were priced in GM shares), then the monetary demand (e.g. monetary value) would dominate demand to hold them as an income producing asset (e.g. “backing value”). Take gold. If Fed tomorrow announces that gold is again legal tender in the country, I imagine the price would skyrocket even though there was no change in its “backing”. On the other hand, if gold suddenly lost all its appeal as a monetary asset and was left only with jewelry and industrial demand, the price would head much lower than is today.

      I agree, there is no reason, why the price after the money supply increase shouldn’t reflect the new backing => 1$ = 0.8 oz of silver + equivalent of 0.2 oz of silver in gold. But that is not the inflation I was pointing at. What will change is the prices of all other goods in the economy. For example If there is only apples than the $ price of apples will rise by 20 % because more $ chasing them. The value of $ expressed in all other goods in economy will fall. With one exception. Exactly as you write, suddenly the same amount of silver buys you less apples, because now the same monetary demand (service) of $ was split between silver and gold.

    4. Juraj:

      The important difference between financial securities (stock, bonds, and bank notes) and goods (gold, apples) is that goods are produced using scarce resources and production functions, so the supply curve of goods slopes up. Financial securities are not produced using scarce resources. They can be issued or retired instantly and costlessly in infinite amounts, so if you were to draw a supply curve of securities (which you shouldn't do, but oh well) then the supply curve would plot as a horizontal line, and the value of the security would be insensitive to changes in the demand for it.

  7. It seems to me the problem here is one in the economics of value (which used to be a hot issue in the 19th century). If you assume that the backing is held in the medium of account, there's no problem, but other assets are subject to risk and uncertainty. After all, deposits in Cyprus and Argentina were money.

    There can be all sorts of money, which can differ in liquidity, risk and domain. It would seem only logical to treat assets in a similar fashion.

    If one bank started issuing paper and buying silver, you have to assume all the other banks would do the same thing, given rational pricing. Moreover, since bank notes are pegged to silver, inevitably notes will be backed by other notes. This creates systemic risk, and soon it all comes to a bad end.

    We just recently saw (several times, in fact) the result of fluctuations in the value of assets relative to liabilities on balance sheets where they differed in kind or duration. I don't believe you can just simplify this problem away.

    1. Anonymous:

      1) Bank notes are risky claims to the bank's assets. Very low risk, but it's there. On the other hand, a coin in my pocket or mattress is also at risk. So banks probably lower the overall probability of losing your money. Also keep in mind that people rarely keep more than maybe 5% of their total wealth in a bank.

      2) Bankers historically dealt with the duration mismatch problem by following the real bills doctrine: Only issuing money in exchange for 1) short term 2) real bills 3) of adequate value. By following rule #3, banks assured that their money would be adequately backed and would not inflate. By following #2 they assured that banks would only issue money when it was needed. That is, that they would provide an elastic currency, that grew and shrunk with the needs of business. Rule #1 prevents maturity mismatching, by assuring, for example, that bank notes are backed by assets maturing in 30-60 days, so that note holders would not have to wait longer than that for payment. The real bills doctrine was developed by practical bankers over centuries of experience. Unfortunately, it was rejected by the academic scribblers of the economics profession.

      I'm always a little surprised when I hear economists worrying about banks "lending long term and borrowing short term". That's not what smart bankers do. If a smart banker has lots of liabilities coming due in 90 days, he holds assets that mature in 90 days.

      3) Backing a note with another note can create what I call inflationary feedback: The dollar loses some backing, so it falls in value. This makes the (dollar-denominated) backing lose more value, which makes the dollar fall still more, etc. I explain the algebra behind this in my paper called "There's No Such Thing as Fiat Money"

    2. Mike, when a bank uses demand deposits to make a 25 year mortgage loan, he is "lending long term and borrowing short term". It may not be what a smart banker does, but it is what real bankers do.

    3. Vincent:
      I hear a lot of people saying that banks borrow short and lend long, but I don't actually know that's what they really do. Do you? A bank that uses demand deposits to make a 25 year mortgage loan might earn extra interest, but over a long enough time span, that bank is certain to get caught in a maturity mismatch squeeze and go broke. A smart banker who makes 25 year loans will finance them with 25-year assets. He'll make less interest, but he'll never be caught in a squeeze and go broke.

      I'll bet there's a finance whiz out there who could prove mathematically that the net present value of the smart banker will exceed the NPV of the dumb banker, once the probability of bankruptcy is properly weighed against the higher interest earnings.

    4. Mike I don't think there are any banks who sell 25 year bonds to raise money for 25 year loans. It would be a very good way to do things.

      Anguilla has 4 banks. Of these 2 are just local, one is Caribbean wide, and one is Canadian. The 2 local banks definitely had trouble with the borrow short term and loan long term. They had demand deposits and CDs that were 6 month, 1 year, 2 year and loans that were 15 years. When people wanted to take out their money they went into receivership.

      In the US the "liquidity crisis" was really the same thing. The difference is that the Fed loaned money to all kinds of banks to get them through.

      There are those of us that think it should be illegal to make loans of longer duration than your deposits.

    5. Vincent:
      "I don't think there are any banks who sell 25 year bonds to raise money for 25 year loans."

      When a bank makes a 25 year loan, it usually sells the mortgage to a pension fund or some other investor, so the bank is only the middleman. The end result is that the 25 year loan is financed by a lender who expects to be paid over 25 years, and the bank is not vulnerable to a maturity squeeze.

  8. Mike, interesting. Is there some way we can put your hypothesis to the test? Is there some combination of future events your hypothesis would rule out, yet another economist, which did not agree with your hypothesis, might think is plausible or even likely? Can we build a list of such events (specific event combinations) and then grade your hypothesis against them as time passes?

    Also, somewhat off topic, but does your view here inform your opinion as to the likelihood of hyperinflation in Japan at all? How likely is it in the next 2 years? 5 years? Is there a reason for your opinion?

    1. Ha, I just realized how incredibly general my question sounds. I can reuse that question on every single economics blog out there! :D (or political science, or just plain science, Lol).

    2. Tom:

      To give a nice general answer, the backing theory says that inflation should be more correlated with the issuer's assets and liabilities, while the quantity theory says inflation should be more correlated with the quantity of (broad) money and real GDP. Those variables are impossibly vague, so empirical testing is unlikely to convince skeptics.

      Just as we can't predict next year's stock price, we can't predict next year's inflation rate.

  9. If the bank has a mix of real things like gold and silver and short term bonds then all can be fine. But if it only has long term bonds there is no real anchor to the currency. It is as if you are backing the currency with the future value of the currency. This does not work well.

    Imagine Japan's central bank only owned 10-year JGBs and nothing else. Then if the Yen is dropping the 10-year bonds drop much more (as the markets project the value at the current rate of drop 10 years out). But then the value of the backing is dropping faster than the Yen, so the Yen will drop even faster and the bonds drop even faster. If they have no other assets, there is nothing to break the cycle.

    Part of "the real bills doctrine" was that bonds should not be long term. If they are short term you can just wait for them to mature and never sell at a loss. But if your currency has dropped in half and the only thing the central bank can do is wait for 10 year bonds to mature, or sell bonds at a huge loss, the reflux is broken.

    Now in practice the Japanese central bank also owns some US bonds. If they have 10% of their backing as US bonds then there is sort of at most a factor of 10 drop before their have enough backing.

    1. An alternative is for the central bank to try to peg the interest rate on long term bonds, so they don't drop in value when measured in Yen. However, if they do this then they have no ability to support the value of the currency if their only asset is long term bonds. Again, reflux is broken.

    2. To be very clear, when the central bank pegs the interest rate at 0.5% for 10-year bonds the way it does this is by buying bonds to support the price of bonds so it stays high enough to give that low rate. But to keep this low rate they have to keep buying bonds. If they have to keep buying bonds, then reflux is broken.

    3. Vincent:

      Suppose a bank has issued $300 of its own notes, and the bank's assets consist of 100 oz of actual silver, plus bonds worth $200 (denominated in dollars). Define E as the exchange value of the dollar (oz/$). Setting assets (100 oz plus bonds worth 200E oz) equal to liabilities (notes worth 300 E oz) yields


      or E=1 oz/$

      So dollars can be partially backed by dollar-denominated assets.

      Now suppose the bank is robbed of 30 oz, which is 10% of its assets. The equation becomes


      or E= .7 oz/$

      The 10% loss of assets has resulted in a 30% inflation due to inflationary feedback. As you said, this feedback effect gets bigger as "real" assets become less, and if there are no real assets, then E is indeterminate.

      The RBD says that BANK NOTES should only be issued in exchange for real bills of adequate value that mature in 60 days or less. That way, the longest the note holders would have to wait for payment is 60 days, and the notes themselves would specify that payment could be delayed 60 days in emergencies. Those old time bankers were thus very smart about matching the maturities of their notes with the maturities of their assets. That's why I doubt the stories about modern bankers being dumb in that respect.

    4. " As you said, this feedback effect gets bigger as "real" assets become less, and if there are no real assets, then E is indeterminate."

      All of this QE the central banks have been doing is buying longer term bonds. This is against the "60 day" rule and I believe far greater than all their other backing at this point. The Fed is like 4+ times the assets and most of the new stuff is longer term. The Japanese are not following RBD either. The modern bankers don't understand the reason for the 60 days or real bills type things. I think we are going to see the "indeterminate" effects.

    5. Vincent:

      The 60 day rule matters much less when the bank notes are inconvertible, like FRN's. Indeterminacy would result from a dwindling of the central bank's real assets, like gold, buildings, etc. It wouldn't result from the Fed's holding of long term bonds.

    6. The larger fraction of the central banks assets that is held in long term bonds the less they have in real things and the more vulnerable they are to a drop in the value of long term bonds starting a feedback loop where the bonds going down causes the currency to go down which causes the bonds to go down etc.

    7. Yes, I see your point about the long term bonds being more sensitive to interest rate changes. But the actual feedback itself is due to the bonds being denominated in dollars, not to their being long term.

    8. A long term bond goes down far more than just the currency or a short term bond. Also, with a 60 day bond you could always just wait and get all your money back plus interest.

      A 10-year $1000 bond paying 0.5% drops to $652 if interest rates go up to 5% and $182 at 20%. It gets even worse for longer term.

      To get a positive feedback loop where things give a really good death spiral you need some amplification. The long term bonds provide that amplification of the downward moves in the currency so the backing really crashes. Wise old bankers knew not to back with long term bonds. There is a reason.

    9. Note that the faster the currency goes down the higher the interest rates the market demands. And the higher the interest rates the lower the value on the long term bonds. The lower the value, the lower the backing. The lower the backing, the more the currency goes down fast. So the backing theory provides a good theory for explaining hyperinflation. It is really my favorite theory for explaining hyperinflation as it always works, where others don''t really cover the case of government stealing the central bank reserves (Argentina style).

    10. I see your point. I had been thinking only of the feedback effect that results from denominating assets in your own money. But there's also that interest rate feedback effect that you mentioned

    11. My focus on hyperinflation as a positive feedback loop that should be explainable in any reasonable theory of money helps in this case. :-)

  10. Unless you want include mortgage, credit card, automobile and student loans and securities derived from them as true bills, they would be a very small part of a modern banks loan assets. Loans, in turn are only part of a modern banks assets, and banks also have earnings from trading, underwriting, off balance sheet entities, all sorts of financial services, and sovereign debt.

    Reliably valuing a bank from its SEC filings is an incredibly difficult job, perhaps an impossible one. So determining the value of a bank's paper isn't just a matter of adding up some numbers and applying a formula.

    Also there is the question of systemic risk if the backing assets are correlated.

    I don't see how a True Bills theory generalized to cover these issues could ever be falsified, which would make it a belief rather than a testable theory when applied to the real economy, regardless of how well it works in simple models.

    But yes, fiat money is definitely worse. It doesn't even make an attempt to be a theory.

  11. Anonymous:

    The original RBD, as practiced and refined by bankers over centuries, had the effect of assuring that bank notes were (1) adequately backed by short term assets (2) elastic, meaning that the money supply grew and shrunk with the needs of business. Bankers rarely held to it precisely. For example, they would often issue notes in exchange for exchequer bills, which clearly did not fit the definition of "real bills", but were nevertheless perfectly good as backing for notes.

    One of the common complaints of note-issuing bankers was that someone would deposit a 1 pound coin on Monday and ask to be issued a 1 pound note in exchange. The very same pound note would reflux to the bank on Tuesday, with the holder demanding a 1 pound coin. So bankers often refused to issue notes for coins. But bankers found that if they issued notes in exchange for real bills, then the notes would stay out in circulation and would not waste the bankers' time.

    Now update that to a world where bank notes have been replaced by checking accounts and credit cards. When those kinds of moneys are issued on Monday, and reflux on Tuesday, it is handled as a mere bookkeeping entry, and the banker's time is not wasted. So nowadays it is much less important for bankers to limit their money issues to people who offer real bills in exchange. Of course, banks should still issue new money only in exchange for short term assets of adequate value, but it's less important for those assets to be 'real'.

    As you said, the RBD is difficult, if not impossible, to test empirically. Then again, it's also difficult if not impossible to test the theory that the value of stocks and bonds is determined by their backing, but everyone accepts that idea.

    1. You can see that all real central banks have backing. Things like MMT and Quantity Theory of money don't see a need for backing, yet it is always there. You can see when something happens to that backing, like the government taking it, that the currency goes down. You can see that to get hyperinflation either something happens to the backing (government theft) or the backing is long term bonds that crash in value as inflation and interest rates go up. So the backing theory fits the experimental evidence really well, better than any other theory I know. And I am not aware of any experimental evidence that really contradicts the backing theory.

    2. Yep. But then read the blogs of quantity theorists like Nick Rowe or Scott Sumner, and they say that they believe the quantity theory because the evidence supports it. And those are almost the only two guys who even talk about the backing theory. The whole rest of the econ profession never even uses the words "backing theory".

    3. Then if we are the only ones armed with the correct theory we should be able to better predict things (like the Yen crashing) and get rich!

    4. Note, with the backing theory both Argentina and Ukraine were obviously headed for trouble. So for me it has made predictions that worked out. I did not invest anything in those though. Trying on the Yen for real.

    5. The other thing about central banks buying long term government debt is that the total debt almost always keeps going up. Really the only way the government pays off one bond is by first selling another. If private people are not buying then either the central bank has to buy a new bond before the government pays off the old one, or the government defaults (does not happen). So basically the reflux breaks and the central bank is just making money and buying bonds from a dead beat that won't pay back. In the end this is bad backing.

    6. Tell those quantity theorist that New Zealand increased their monetary base by a factor of 4 and never got inflation even years after.

      Was JP Koning that found this as part of a contest I did. See comments at:

      It makes perfect sense from a backing theory point of view. They made New Zealand money and bought Euros. Then later they used the Euros to buy back their New Zealand dollars. A quantity theory would have expected high inflation, a backing theory would not. Backing theory was right.

    7. "if we are the only ones armed with the correct theory we should be able to better predict things (like the Yen crashing) and get rich!"

      That reminds me of a talk given by Steve Givot, who had been the research assistant for Black and Scholes when they developed their option pricing formula. Givot started trading on the CBOE, and he said that for the first 6 months he lost money. When he talked to other traders and asked how long it took for them to start making money, they answered "about 6 months". It makes me suspect that the same thing would happen to anyone who tried to profit with the backing theory. Put another way, currency traders probably already are using the backing theory whether they know it or not.

      Oh yeah, and I also chicken out whenever large amounts of my money are at risk.

    8. Yes, probably currency traders that make it have figured out what is important.

      Well, if the Yen crashes and you see me building a good sized yacht on then my experiment turned out ok. :-)

  12. When I look at the media reporting regaring currencies it is clear that a very broad range of factors are weighed in. So at least short term the mainstreams opinion is that currencies are depending on the strength of the economy. That feels very close to the backing theory. There could be an argument made for a dependency chain of central bank/government/taxes/strength of the economy. With that said Japan has a huge amount of productive capacity and relatively low taxation. They reason currency traders hasn't bid down the yen yet must be that despite all the debt of the central gov and the balance sheet of the BOJ they still see untapped potential further up in the dependency chain.

    1. Good point. But hopefully Vincent's Yacht plans won't be affected!

  13. Some random thoughts:

    If you normalize currencies by their purchasing power parity, then rational pricing implies that the resulting differences should represent the price of the additional services the currency provides.

    This allows value to come in part from fiat (like need for certain forms of transactions) or almost entirely from liquidity services, but only to the extent that that arbitrage permits.

    Inflation requires arbitrage opportunities, since it is a matter of relative prices.

    If the value of backing is based on other backing, then it is possible to have backing instabilities from systemic effects and revulsion. This certainly seems to be what we actually saw with subprime mortgage backed securities.

  14. "This allows value to come in part from fiat..."

    JP and I have gone back and forth about that quite a bit, and I usually end up caving in and admitting that maybe there can be a premium of 5% or so above backing. But then look at real moneys issued by real banks, and we see that if a bank has assets worth 100 oz of silver, backing $100, then $1=1 premium. Also remember that it's easy to imagine a money premium for GOODS like gold and silver, but the monetary premium seems very unlikely for financial securities--things that can be produced or retired instantly and costlessly in infinite amounts.

  15. "If the world has too many spoons, then people will find it profitable to melt silver spoons."

    Why? If it is profitable to melt silver spoons, then it will be profitable for someone with an extra silver spoon to sell it to someone who wants to melt it. That is, the spoon will be worth less on the market than it was before there were too many. Given that, how do we know that it will be profitable to melt them? In the early 1930s mid-western farmers grew too much wheat. They left piles of it by the side of the road. No profit there.

    There is also a problem with only having spoons and bullion in the picture. If people came to prefer bullion over spoons, they could simply melt their own spoons, without selling them. Where is the profit in that? For them to make a profit, there would have to be another use for the bullion. Maybe there would be a demand for silver bullets to kill vampires. Then silver spoons might be melted down to make silver bullets and sell them for a profit.