Sunday, February 23, 2014

When money ceases to be an IOU

Despite its odd denomination, the above is a genuine banknote. Ne Win, Burma's military dictator from 1962 to 1988 and the man pictured in the note, was convinced that his lucky number was nine. So he had notes issued in multiples of nine, including the forty-five and ninety kyat notes (which also each add up to nine 4+5=9, 0+9=9). But I'll get to this story later.

This is a continuation of a post I wrote earlier describing how central banknotes aren't mere bits of intrinsically useless paper. The standard view among economists is that banknotes are bubble assets. Because they are intrinsically valueless, the fact that paper notes earn a positive value can only be explained by the fact that the market expects them to have value in the future, much in the way that a ponzi scheme or chain letter is perpetuated. This view goes back to Paul Samuelson (pdf), who described how a "grand consensus" might be arrived at whereby society could contrive to have "oblongs of paper" pass at a positive value from generation to generation. The creation of such a bubble would allow for an efficient allocation of resources. As long as each generation could fool itself into thinking that the next would accept these worthless bits of paper, the scheme could continue indefinitely.

However, if real life banknotes were in fact pure Samuelsonian bubbles, then no one would bother participating in central bank redenominations. If a central bank were to announce a 10:1 redenomination, why take in your $1000 note to be redeemed for a $100 note? After all, your existing $1000 note will buy you more stuff than a prospective $100. That we do take part in redenominations is due to the fact that the central bank is threatening to do something to its legacy note issue—it is threatening to cease honouring them as its liability, or IOU, an act sufficient to drive that notes' value to zero. This quality of a banknote as a liability or IOU means that a note is not a Samuelsonian asset.

That central banks offer legacy note conversion at all  is another sign of the IOU-nature of bank notes. It explains why we don't see revolutions during redenominations and demonetizations. Those suddenly left holding old paper would riot on the streets upon the sudden displacement of their notes by a new currency. We don't see riots due to redenominations or demonetizations because banknotes aren't Samuelsonian—central bank paper carries an implicit IOU whereby the central banker will typically repurchase the legacy notes that they've issued at a fixed rate rather than letting those notes flap around in the wind.

For instance, when the European central banks ceased their issuance of domestic banknotes they didn't strand them. Rather, they offered a window in which individuals would convert legacy notes into euros. This window varied in size by country. In the case of the Banca d'Italia, for instance, lira could be converted into euros until February 2012, a ten-year window. De Nederlandsche Bank continues to offer to convert old Dutch guilders into euro, and will do so until 2032, while the Bundesbank, Banca D'Espana, and the National Bank of Belgium guarantee to perpetually redeem deutsche marks, pesetas, and francs respectively with euros. Here is a full list:

As a convertibility window narrows toward zero, a central bank's paper loses its IOU nature and approaches the ideal of Samuelsonian money. For instance, when Saddam Hussein decided to cancel the old "Swiss Dinar", he did so by offering note holders a tiny six-day exchange period, a mere crack compared to the massive windows left open by European central banks. Saddam narrowed this aperture even further by deviously shutting down the Iraqi border during the exchange period, preventing the large population of merchants in Jordan and elsewhere from crossing over into Iraq to swap their Swiss dinars for new dinars.

The Burmese kyat, pictured above, is another example of a note that approaches bubble money. In September 1987, Ne Win demonetized the largest denominations of kyat: the twenty-five, thirty-five, and seventy-five kyat notes, issuing in their place the aforementioned forty-five and ninety notes.

As dictator, Ne Win didn't seem to see any problem in burdening the Burmese population with these awkwardly-denominated notes. The legacy notes being replaced that September were already in odd denominations (25, 35, and 75), the hallmark of a redenomination in 1985 that replaced more conventional denominations of twenty, fifty, and one-hundred kyat notes. It seems that the seventy-five note was issued to mark Ne Win's 75th birthday, and the thirty-five replaced the fifty because it was, according to Ne Win's numerologist, the luckier of the two numbers.

To make matters worse, the legacy issue of kyat notes could not be converted into new forty-five and ninety kyat notes, effectively removing the IOU embedded in kyat notes and rendering a large part of the nation's stock of notes worthless. Here is an account of a Burmese individual at the time:
Then on September 5, without alerting even his cabinet, Ne Win passed a sealed envelope to the Information Minister. Inside were instructions to broadcast the enclosed announcement at 11:00 a.m.
I listened to the 1:00 p.m. news on the car radio outside the art exhibition. When the announcer said government employees would receive 450 kyats as emergency funding, I knew our kyat notes in denominations of 75, 35, and 25 had all been instantly devalued. At home that afternoon my family members gathered to pool our currency. Ko Gyi had received cash payment from a client on Thursday, all in 75-kyat notes, and I had not yet opened my two most recent pay envelopes, one containing my salary, the other my bonus, also dispensed in 75-kyat notes. My brother and I together had only 25 kyats in 10- and 5-kyat notes. Our other cash was useless.
Still hoping from compensation after the demonetization announcement in 1987, some people hoped to make a quick profit. Expecting a grace period, they started collecting 75 kyats by paying 50, then collecting 30 kyats for the same amount, then 20, until finally, when no follow up news came from the government, they realized they had lost everything. Two thirds of the total currency in circulation became useless paper, and ordinary citizens lost their entire savings. Even the beggars suffered.  - No Time for Dreams: Living in Burma Under Military Rule, by San San Tin
Ne Win had contrived to create Sameulsonian money.  Rather than greeting Ne Win's contrivance with open arms, popular uprisings began in 1988, eventually leading to Ne Win's resignation. Aung San Suu Kyi would emerge as an opposition leader in these uprisings.

North Korean won are a more recent example of notes devolving into mere "oblongs" of paper. In November 2009, the North Korean government announced a 100:1 redenomination. Nothing unusual here, except that North Koreans were given a seven-day window to convert existing Korean won into new notes. This window, already narrow, was further constricted by a ₩100,000 limit on the amount that could be converted. Given the then prevailing black market exchange rate of ₩3,000 to the dollar, the amount of legacy won that could be converted into new won was limited to around $30. Tragically, any North Korean who had won-denominated savings larger than that amount lost much of what they had.

Public anger forced the government to increase the limit to ₩150,000, but widespread uprisings as in the case of Burma never followed. The perception that future surprise redenominations might rob existing note holders has not promoted the stability of the won. Indeed, the inflation rate in North Korea is one of the highest in the world. Circulation of US dollars and Chinese yuan increased dramatically as a result of the redenomination, since no one fooled once wants to be fooled twice. (A shift to the use of the dollar also occurred in Burma as a result of Ne Win`s 1987 redenomination, as this interesting article points out.)

Though Samuelsonian money is an elegant idea, the reality is ugly. Only the most sinister dictators have tried to perpetuate it. The contrived currencies that have had Samuelsonian characteristics—Burmese kyat, North Korean won, and Saddam dinars—have been among the world's worst currencies. Widely accepted and long-lasting bank notes like dollars and euros are not bubble-like. Their issuers threat these as IOUs, much like a stock or a bond. If they fail to take these commitments seriously, say in the case of redenominations, their currencies will quickly be out-competed by currencies issued by banks that do.


  1. "To make matters worse, the legacy issue of kyat notes could not be converted into new forty-five and ninety kyat notes, effectively removing the IOU embedded in kyat notes and rendering a large part of the nation's stock of notes worthless"

    They demonetized the kyat as state currency and replaced with a different currency. That is not removing the IOU at all. They purposely killed the currency. Its not the same as continuing to use the same currency with the IOU aspect removed and replaced with equity.

    "Though Samuelsonian money is an elegant idea, the reality is ugly. Only the most sinister dictators have tried to perpetuate it. The contrived currencies that have had Samuelsonian characteristics—Burmese kyat, North Korean won, and Saddam dinars—have been among the world's worst currencies. Widely accepted and long-lasting bank notes like dollars and euros are not bubble-like. Their issuers threat these as IOUs, much like a stock or a bond."

    Stocks aren't IOU's, they are equity.

    1. The 1987 redenomination didn't replace kyat. The currency continued to exist, as did the issuer, the Union Bank of Burma. But the 25, 35, and 75 notes ceased to exist as IOUs.

    2. "As a convertibility window narrows toward zero, a central bank's paper loses its IOU nature and approaches the ideal of Samuelsonian money."

      It ceases to be the official accepted medium of exchange altogether or is demonetized. Thats not the same as removing the IOU nature of a currency while it is still the state accepted medium of exchange.

      The 25, 35 and 75 notes were demonetized altogether which is entirely different.

      Issuers dont need to treat a financial asset as IOUs. An example of this is stocks which are treated as equity.

  2. Is it illegal to produce counterfeit copies of the old money?

    Can you get the right relative supplies of different denominations of the old money?

    Did people obey Ne Win because each person expected everyone else to obey him? (Are governments bubbles?)

  3. Suppose you have two equilibria: one with green money and one with yellow money. A sunspot could flip you from the green equilibrium to the yellow equilibrium. Or, if one large merchant, who everybody trades with, decides to switch from using green money to using yellow money, that could cause the flip. The government is a large merchant.

    1. I'm not disputing the theory behind green and yellow money, but the evidence. Why don't we have more Samuelsonian green and yellow paper bubbles? USD, C$, and Euros aren't ponzi schemes, they're IOUs, while the only notes that appear to approach the Sameulsonian ideal are the worst ones, like kyat and NK won and whatnot.

    2. Even if a large merchant continued to accept demonetized notes, banks could no longer take those notes to be deposited at the central banks, and worn notes could not be exchanged for crisp notes. Good notes would be hoarded and bad notes would circulate even faster, these notes earning ever larger discounts. At the same time, counterfeiters would compete to issue copycats. Eventually the value of demonetized notes would fall to the cost of printing, like in Somalia.

  4. JP, a brief off topic survey question for you here. I think I put it best to David Glasner:
    Your thoughts?

    1. My first thought is that it would be irrelevant, like a stock split.

    2. JP thanks much. My first thought too, but I didn't think like a stock split. Mark Sadowski came along and kindly discussed the matter with me at some length, and ... well, I'm still not 100% satisfied. I summarize my three sequentially obtained incompatible (and wrong?) conclusions today in this brief list:

      If it seems like I'm going to extremes to find an absurdity... well, that's fair. But I think that's a legitimate exercise. It works for math. I'm not to the point where I feel it really all hangs together yet here though. I'm still not getting the "magic" of base money... or you could say that the fact that it still seems to me to require magic is a problem.

  5. I agree that bank notes are not bubble assets but I don't think this argument goes far enough to reach that conclusion. If they are IOUs then to what exactly do they entitle the holder? The argument seems to be that if the CB decides to switch to another type of note that they entitle the holder to some quantity of those notes but if you adhere to the bubble explanation then don't you argue that this is just a bubble asset backed by another bubble asset?

    This seems to amount to the claim that money is in fact a bubble but that somehow the CB has the ability to prop the bubble up indefinitely by promising to prop it up indefinitely. This is an identical argument to the Samuelsonian argument except that it relies on some mysterious power of CBs, without which the asset would lose its value. From whence does that power fundamentally derive?

    1. Mike, most of my arguments about the IOU nature of banknotes are in this post. The central bank's promise to provide new currency to demonetized noteholders is only one portion of the overall IOU that a central bank takes upon itself.

    2. I had read that post and I still think it misses the heart of the matter (though it is on the right track). My complaint is this:

      "If you are unwinding the central bank, you are unwinding the dollar itself. So let’s imagine that the CB has some amount of gold in its vaults and it has some amount of notes outstanding which are denominated in dollars. Those dollars are said to be a senior claim on the assets of the bank (the gold) and let us assume that there are some other claims (these may be dollar denominated debt or undenominated equity). But the notes do not entitle the bearers to a specific quantity of gold. So how much of those assets are they entitled to?

      In this case the assets cannot be converted into dollar terms and divvied up in the same way as with an individual bank because the value of a dollar is not clear. The whole market for dollars which forms the basis for the valuation of the assets relative to the various claims in the former case is the very thing being unwound in the latter case. So how much of the gold goes to note holders versus other claimants? The answer is not at all clear. So it is hard to see how this claim serves as any kind of foundation for determining the value of a dollar."

      This from a longer response which I think also solves that problem and can be found here:

    3. Good point. It would be very challenging for a bankruptcy judge to properly reward each claimant upon a winding down of a central bank. I don't have any great responses.

      One approach would be to use the gold price just prior to the winding down announcement as a reference point. If gold was trading at $1300, the purcasing power of $1 would have been around 22 milligrams of AU. In the winding down, all noteholders would be entitled to 22 mg, and once they have been made whole, then bond holders and other lenders would get their share, followed by shareholders who get the rest.

      Until then, the main determinate of the value of the dollar is the central bank's promise to maintain price stability, which it enforces by selling "backing" assets should the dollar fall to fast.

    4. I did a little "back of the envelope" calculation and it seems they would need more than 2000 times as much gold as what they officially have at Fort Knox (more gold than has ever been refined in the history of mankind) to redeem the current monetary base at $1300 per oz. I don't think your theory is completely crazy, it seems like all you are really saying is that people think the CB will always be willing to give them something for their dollars and what that thing is doesn't really matter.

      But even if we assume that the Fed has as much gold as they claim (which some people seem to be skeptical about) and that in the event of a currency collapse they would simply take what they have and divide it by the money base and hand it out and we ignore the issue of what to do with all the M1, M2, etc. that makes up most of what people consider their "money" we are left with an implied backing that amounts to a tiny fraction of the current real value. This means that you still have to explain the rest of the value. One explanation could be that it is a bubble built on a small real foundation like a pearl built on a grain of sand. I think that is a bad explanation. I think my explanation is much better. I'm trying to get you to read it but I don't want to be a troll so I will probably leave you alone if this last attempt fails. =)

    5. So according to this:
      There's 5.6 billion troy ounces total ever mined = 174179470 kilograms. Dividing by 22 mg / $, we get 7.9172486e+12 or about $7.9 trillion, right? That's what the world's stock of gold is worth?

      Fort Knox has 4578000 kilograms according to this (if I read that right):
      or 1/38th of the total amount ever mined? Wow, that's more than I would have thought. So about $207 billion worth?

      So Mike you say there's more than 2000 times as many dollars out there? That would be more than $414 trillion dollars... is that right??

      Isn't a more reasonable calculation just the total base money... reserves are now the major component and they are at $2.5 trillion, right? Now how many bonds does Treasury hold? According to Mike Sproul you subtract those from both sides because they're both denominated in $, before dividing remaining Fed assets by remaining outstanding $, to get assets/$.

      ... I don't have the patience to carry this through, but we're getting back into more reasonable territory here, don't you think? How much gold does the Fed own? Any?

    6. BTW, I love using google as a calculator for this stuff: Want to know how many teaspoons in a cubic light year: no problem, just type it in google:

      1.71788573 × 10^53 teaspoons

      So how many moon masses in an Earth mass?: 81.2800178

    7. I suppose the reciprocal is more "useful" ... ' add 5.8e-54 cubic light years of sugar'... :D

    8. haha. Here is where I got my number for the money base.

      It seems to be about 3.8T. And got 147.2 million oz. troy "in Fort Knox" from here.

      That makes about $191B of gold at $1300/oz. So I think I added an extra 0 (that's what happens when I do calculations on the back of an envelope). My basic point remains though, it's not enough gold to cash in all that money at anything like the current market value.

      As far as subtracting the bonds, this essentially gets at my main point which is that it is the nominal debt (and assets "backing" that debt) which "backs" the dollar but once you start thinking about it this way, you start to see things differently. The real assets on the CB's balance sheet are not at the crux of the matter.

    9. Mike, I read your blog post yesterday. I'll try and leave a comment on it this weekend.

      There's a difference between an orderly windup and and a disorderly windup of (say) the dollar. In an orderly windup, people would be notified ahead of time of the Fed's coming closing, in the same way that a retiring businessman might notify shareholders that he was planning to shutter a successful business and cancel all shares.

      To avoid chaos, all existing dollar-denominated contracts would have to be re-denominated into some other successor unit, say gold or yen or x. This would include dollar-denominated bonds, which would now pay out gold or yen or x. The Fed would now have its existing gold ounces plus its newly-redenominated and still valuable bond portfolio which it could in turn sell for ounces, and then proceed to cancel dollar bills one-by-one.

      But I'm not confident about this. As I pointed out, I still think the key IOU underpinning the value of modern dollars is the price stability promise.

      (Incidentally Tom and Mike, be careful of the Fed's gold holdings).

    10. Mike, your "private debt chartalism" story[1] is the best answer I've seen to the Williamson-Noahapinion "fiat money = bubble?" blogapalooza of 2012. I also don't think it's incompatible with the Nick Rowe-JP "CPI stability promise" idea (although I can't see how that alone could underpin an inconvertible currency's value). Rather, debt chartalism (the need to extinguish dollar-denominated debts or else lose your house/car) is the actual "thing" which gives fiat money value; the central bank keeps the system from spiraling out of control by steering the growth of NGDP/fall of the $'s value in CPI-units (required, as you further noted, due to bank interest) in a predictable, orderly way to keep hyperinflation or deflation--the only potential threats--in check.

      I also don't see how the wind-up value of the central bank's assets (Sproulian backing) much influences the dollar's day-to-day value. But I think it does have a function. I've asked a few bloggers (such as Glasner, who has repeatedly called gold an irrational bubble) why central banks continue to hold so much of it (around a fifth of global stocks) or continue to add to their stocks. I didn't really get a satisfying answer[2]. My own (not unique) take: a fail-safe device. In the worst case scenario, central banks can still pin down the value of fiat currency by making it convertible into gold. My question: if all central banks took Glasner's view of gold's irrelevance and sold off their holdings, would the value of fiat money remain unaffected? If the answer is no, can we truly say that we've gone off the gold standard?


      Mike, you said that you didn't think that anyone else has told a story quite like yours about fiat money. I certainly haven't seen one, but JP--seemingly familiar with every monetary idea under the sun :)--might feel differently. So has Mike added a new wrinkle, or did someone say essentially the same thing in 1836 (or thereabouts)?


    11. it's not enough gold to cash in all that money at anything like the current market value.

      Mike, keeping in mind the caveats mentioned by JP in his post, the US govt (i.e. US Treasury) owns around 261 million troy ounces.

      So at $1300 an ounce, that's around $340 billion. But who says that the govt/central bank has to redeem 100% of the currency in existence to restore the gold standard (i.e. link the value of the dollar to some real asset)? Can't the central bank just set the gold reserve ratio at some appropriate % of M1? Larry White makes the case for gold (which I don't necessarily support) here:

      "Does the US Treasury own enough gold to return to a gold-redeemable dollar at the current price of gold? Yes, assuming that they have what they say they do. At a market price of $1600 per fine Troy oz. (to choose a recently realized round number) the US government’s 261.5m ounces of gold are worth $418.4b. Current required bank reserves are only $83b. Looked at another way, $418.4b is 19.9 percent of current M1 (the sum of currency and checking account balances), a more than healthy reserve ratio by historical standards.[6] Combined with the likelihood that US citizens’ hedging demand for gold will shrink by more than banks’ reserve demand will grow with larger real demand for M1 balances, I expect that the denationalization and remonetization of the US bullion stock at the current price would allow the US economy to export some excess gold. There will be a small transitional windfall for US citizens, getting imported goods and services in exchange for excess gold."

      Incidentally, I searched your website for "Selgin" and didn't get any hits. :( Any general thoughts on Selgin/White's free banking research (or Selgin's "good deflation/productivity norm), or Glasner's free banking book?

    12. Just a few thoughts.

      I like John's description of your idea as private debt chartalism. Bits of paper can earn value intrinsically or extrinsically. Backing is intrinsic while chartalism is extrinsic. In the case of chartalism some external body, say the the sovereign, sets a liability that can only be extinguised by proferring a certain piece of paper. Your debt story is the same, except the external institution that requires settlement with notes is the private debt market.

      I wrote a post a while back on how the requirement to settle debts (legal tender laws) might give bits of paper some positive value. In theory it makes sense, but I don't think that the value of modern banknotes are actually determined in this way. I do think that the requirement to settle debts with banknotes does add a liquidity premium to the value of modern banknotes.

      Some criticisms:

      1) So why do we observe central banks holding assets? If the value of banknotes is maintained by the demand to settle debts, then it would be redundant to own assets. Your theory would seem to imply that the world's 170+ central banks are acting irrationally.

      2) What anchors the price level? Why does $1 buy one apple and not $10 apples? Does a Canadian dollar by 0.98 US dollars because there are less debts to settle in Canada?

      3) How do you explain inflation? I can explain it by saying that the central bank targets a 2% rate of decline in banknotes, and that it reduces the backing of notes. Your route will be much harder. Does a currency inflate because there are fewer debts that must be settled? Does it deflate because more debts come into existence?

      4) Why are deutsche marks still valuable? There are no longer any mark-denominated debts, so according to your theory, marks should be worthless, or close to it. Yet if I offered to unburden someone of their deutsche marks without offering anything in return, they'd laugh at me. That they remain an IOU is the most likely explanation for this.

      5) How do you explain the debut of a new currency? Perhaps the demand to settle debts with a certain brand of note would give an initial value to those notes, but why would anyone denominate debts in a currency that as-yet has no value? There's a circularity problem here. There's no way to kickstart a new banknote into existence if it doesn't already have some positive value.

      [Cross posted at]

  6. P.S. a stock (equity) is an IOU it just isn't nominally denominated. It entitles the bearer to x% of the assets/profit of a company.

  7. JP, I pointed out this article to hyperinflationist Vincent Cate: it inspired him to add another reason for hyperinflation to the list he's collecting. :D
    "Fiat money is a bubble, hyperinflation is when the bubble pops."

  8. John S

    Thanks, I think you have successfully reiterated my basic thinking on this which is reassuring. As far as winding up the CB/returning to a gold standard, I actually do think they could return to a gold standard, I actually think they have plenty of gold to do this but only because of all the debt outstanding which you could use to cancel out positive dollar balances. People have recognized this in discussing this issue but seem to then forget about it when they explain the value of a dollar. It strikes me as an "dark matter" situation where there is this model of the monetary universe but it doesn't really make sense with the quantity of matter (money/gold) that we actually observe. In order for it to work, there has to be something else out there. I think that thing is debt and once you start looking for it, it is all around us and plain as day.

  9. Also, I am a fan of the Free Banking crowd but I also think Selgin et. al. are not noticing the importance of debt in determining the value of a dollar (and a wide range of implications of this).

  10. Here is a follow-up. Doesn't answer anything but hopefully it helps explain how I see the relationship between fiat money and money backed (explicitly) by hard assets. I will get into monetary policy and stuff next but some of the implications should start to become clear if I am managing to make myself clear here.

  11. John S,

    P.S. I dug up this post where I mentioned free banking and said I liked it but I never really got into discussing it haha.

    I'm not an expert on their work but I do read their blog now and then and I like that they are against central banking (more or less) without being against fractional reserve banking (like many Austrians). The Rothbardian argument against fractional reserve banking is silly in my opinion (as I mention briefly in the post above).