Thursday, January 23, 2014

Banknotes: IOUs or not?

The building that housed the original Swedish Riksbank, the earliest European issuer of banknotes

In my previous post on bitcoin, I noted offhandedly that I don't consider modern central bank-issued banknotes to be fiat but classify them as liabilities of the issuing bank. By fiat, I mean unbacked, intrinsically valueless, inconvertible, unenforced bits of paper. In the comments section, monetary operations whiz ATR (who blogs here) challenged me to clarify the nature of the IOU attached to paper currency, and whether that IOU had any value.

It would be cheating to point out that the world's 160+ central banks all list banknotes as a liability on their balance sheet. The deeper reason that I prefer to classify banknotes as liabilities (i.e. IOUs, promises, claims, or obligations) rather than fiat bits of paper is the "fine print". The best place to find the clauses governing the IOU nature of banknotes isn't on their face, but rather in the various acts and legal documents that govern a given central bank.

Section 16.4 of the Federal Reserve Act, for instance, specifies that:
Federal Reserve notes issued to any such bank shall, upon delivery... become a first and paramount lien on all the assets of such bank.
In the notes to the 2012 audited financial statements of the Federal Reserve system, section 3(k), it says:
To satisfy the obligation to provide sufficient collateral for outstanding Federal Reserve notes, the Reserve Banks have entered into an agreement that provides for certain assets of the Reserve Banks to be jointly pledged as collateral for the Federal Reserve notes issued to all Reserve Banks. In the event that this collateral is insufficient, the Federal Reserve Act provides that Federal Reserve notes become a first and paramount lien on all the assets of the Reserve Banks.
If we look at the Bank of Canada Act, Section 25 (1),
The Bank has the sole right to issue notes and those notes shall be a first charge on the assets of the Bank. 
Section 34.1 of the same Act says that:
...in no case shall the affairs of the Bank be wound up unless Parliament so provides, but if provision is made for winding up the Bank the notes of the Bank outstanding are the first charge on the assets.
So in the case of the two central banks that I'm most familiar with, banknotes are ultimately claims on whatever stuff the central bank happens to have in its vaults. This means that on the occasion of the winding down of the Fed or the BoC, note holders are entitled to receive real assets, in the same way that a bond holder or stock holder would have a claim on a company's property, plant, & other assets upon the dissolution of that company. Banknote holders have an added bonus of being senior to other claimants, since notes provide a "first claim" in the case of the BoC, and a "first and paramount lien" in the case of the Fed.

Now the idea of shutting down either of these central banks sounds like either science fiction or a free banker's wet dream. What about the long period between then and now? Here are a few other indications of the IOU nature of banknotes. When a central bank chops some zeros off of its existing inflating notes, it doesn't leave its old issue stranded. Holders are entitled to bring their legacy notes in for conversion. Nor do central banks try to wiggle out of their obligation to redeem mutilated currency. When a note gets so worn that it becomes unusable, the central bank will replace it with a crisp new note. Lastly, when central banks merge (as in the case of the Euro) or are succeeded by another central bank (e.g. the Central Bank of Russia replacing Gosbank), holders of orphaned notes are given a window during which they can convert into new currency.

If central banks had no liability whatsoever for their note issue, then redenominations, mutilations, and successions would result in large capital losses to stranded note owners. That they don't would seem to indicate the opposite.

However, in the long waiting period until ultimate dissolution, the most important interim IOU provided by central banks is the promise of stability. The corollary of this in the old days was the obligation to redeem notes with some fixed amount of gold, or, during Bretton Woods, some quantity of dollars. Nowadays, central banks promise to ensure that the value of notes doesn't fall more than x% against a CPI basket. They guarantee to buy back IOUs in a sufficient quantity with the assets held in their vaults if the x% limit is exceeded, to the extent that a central banker might conceivably buy back and cancel every single IOU to enforce their promise.

Some central banks like the Reserve Bank of New Zealand encode this promise into their constituting act, along with stipulated penalties in case of non compliance. Other central banks provide this guarantee more informally, either in the form of accepted practice, tradition, or repeated verbal promises. Take the ECB, for instance, which uses the small quota of characters available on its Twitter account to state that its "main task is to maintain the euro's purchasing power."

Are these good liabilities? Not really. The promise of a payout upon central bank dissolution is a distant promise. In the meantime, the obligation to maintain price stability is subject to all sorts of political pressures to weaken the price target, as well as being held hostage to the skills of the central banker in hitting his/her target. Other central bank policy goals including employment and growth targets may interfere with the sanctity of the price stability IOU. The dubious nature of central bank IOUs isn't a new phenomena. Even in the gold standard/Bretton Woods days banknotes were constantly being devalued or rendered inconvertible.

Bad IOUs central banknotes may be, but IOUs they are nonetheless.

40 comments:

  1. Thanks for the shout-out, but I don't consider myself a 'whiz.' Simply doing my best to understand the literature that's out there, which I've only recently begun to do :).

    "It would be cheating to point out that the world's 160+ central banks all list banknotes as a liability on their balance sheet."

    Great - I'm glad we agree on this.

    Nice find and good points on all the lien stuff. That does seem to be a way to attach some sort of value to fiat, but in an extremely uncertain and discounted (as in time value of money) manner. It's your last 3 three paragraphs that I think get to what we're all wondering about.

    If I recall correctly from the previous post, your hypothesis is fiat money can escape the circularity problem if it has value as an IOU. This is similar but stands apart from the Austrian theory you described, which requires the money to have had pre-fiat value through some sort of fixed exchange rate. So would you agree the ultimate test of your hypothesis would be to start a new fiat currency (initially valueless), have a central bank promise price stability, and see if the money establishes value? It might be worthwhile to at least run the thought experiment. (Or perhaps you could come up with a different test.)

    ReplyDelete
    Replies
    1. "So would you agree the ultimate test of your hypothesis would be to start a new fiat currency (initially valueless), have a central bank promise price stability, and see if the money establishes value?"

      I think that would work. But in order to set an initial positive price for your currency and promise price stability thereafter, you would need to define the currency in terms of some other good or basket of goods, or some sort of financial asset. Otherwise I don't know how the market would know what price to set on the currency, or how the central banker could possibly know how to keep the currency stable.

      Delete
    2. Right. Say the CB says $1 will always equal 1 apple. Is it easy to see why people would start using the currency, and how the CB would certainly be able to defend this?

      Delete
    3. Yep. It's like any IOU. An empty promise to provide something won't have value. The IOU has to be defined vis a vis something else, like an apple, in order to debut. (Unless you can "paint the tape", using confederates to kickstart it with a positive value, as in the bitcoin post)

      Delete
  2. Gift cards, unused prepaid phone credits, Canadian Tire money are all liabilities on the balance sheet of their issuers, but none are claimable in bankruptcy.

    ReplyDelete
    Replies
    1. [Sorry, didn't finish my thought. Continuing ...]
      So my perception of their value is not whether an issuer lists it as a liability, but whether, I can redeem those liabilities for services or sell them. I once held a gift card of a company rumured to be filing bankruptcy, and spent it the next day.

      Delete
    2. Well you're in luck with your US and/or Canadian banknotes, since unlike your gift card they would have a legal claim on assets should either central bank go into bankruptcy. At least so sayeth the fine print.

      Delete
  3. Another great post JP. Thanks. I'm very glad you did this one, because I always draw the personal IOU / bank note analogy, but always feel a wee bit guilty when I do... now I'll have a link to provide for justification... and I like how you snuck in the bit about 160+ central banks while declaring it was unfair to do so. But who's counting, right?

    I hope this attracts the attention of Nick Rowe because I suspect he'll complain.

    ReplyDelete
    Replies
    1. JP's posts *always* attract my attention.

      "Nowadays, central banks promise to ensure that the value of notes doesn't fall more than x% against a CPI basket."

      That, to me, is the only sense in which Bank of Canada notes are a liability.

      And they are a negative real interest rate liability.

      Most of the assets of the BoC are themselves IOUs for banknotes. If the banknotes were worthless, so would be those assets.

      Delete
    2. Nick, though I favor the IOU view, your comment about the assets all essentially being IOUs for more banknotes is exactly the point I was hoping someone would make. That one always raises a question in my mind too. Now all we need is Mike Sproul to drop by and put in his $0.02.

      Delete
    3. Thanks Nick & Tom.

      "Most of the assets of the BoC are themselves IOUs for banknotes. If the banknotes were worthless, so would be those assets."

      I think I agree with that. Hopefully the BoC holds a few inflation protected bonds in its vaults -- that would prevent disaster.

      "And they are a negative real interest rate liability."

      ...with a high enough liquidity return that people are just willing to hold them?

      Delete
    4. JP: "..with a high enough liquidity return that people are just willing to hold them?"

      Yes. But it means that, unlike other liabilities, central bank money is net wealth. That's where the Pigou effect comes from.

      Delete
    5. Are you saying that the liquidity return represents net wealth?

      I'd argue that bank notes aren't necessarily net wealth since, much like government bonds, the public realizes that ultimately they will be taxed to uphold their value. (That's assuming that Ricardian Equivalence holds).

      That capitalized portion of a bank note's value that is due to the liquidity return, however, is net wealth to the community. A liquidity return is a consumption good--unlike the IOU component of a banknote, the liquidity portion was not created in a pair that sums to 0.

      I'd also argue that other goods & assets apart from central bank are net wealth, since liquidity returns aren't isolated to banknotes. I think Pesek and Savings made this argument a while back? (In my spare time, I've tried to follow that debate... Pesek & Savings, Johnson, Friedman, Shaw... it's very interesting).

      Delete
    6. JP. Yep. Pesek and Saving basically had it right, I think.

      Delete
    7. "Most of the assets of the BoC are themselves IOUs for banknotes. If the banknotes were worthless, so would be those assets."

      If the loans are repaid, then the bank notes disappear, and there's no problem. If the loans default, then the bank gets whatever collateral has been pledged. So bank notes can't become worthless unless the collateral is worthless.

      In the case of a central bank owning government bonds, it's doubly indirect, because government bonds are backed by another nominal liability, taxes. But taxes are backed by the ability of the government to seize the real assets of non-payers. So despite the multiple levels of indirection, there are real assets backing money.

      Delete
    8. I fully agree with JP post. I think that CB banknotes are liabilities in the full accounting sense, as there is the possibility that they could result in an outflow of economic resources from the entity issuing those banknotes (i.e. a CB debtor can use the banknote to pay a debt wihin the CB's assets).

      "But taxes are backed by the ability of the government to seize the real assets of non-payers"

      And taxes are generally defined as a % of real production, not in nominal fiat units.

      Delete
  4. "the most important interim IOU provided by central banks is the promise of stability"

    I'll be a minority of one and disagree with that.

    That's an undertaking, but hardly a liability.

    There's no certainty that such a promise can be fulfilled. So it really can't be a liability.

    The true liability is the obligation to exchange for something of equivalent nominal value - such as happens in the recurring two-way exchange of notes for banks reserves or of new notes for old notes, etc. Or for other assets in the event of "wind-up", whatever that involves.

    The balance sheet liability is what it is in legal terms.

    I think the price level stuff is a separate issue, but its not a liability, because there's really no further direct consequence for the central bank - in nominal balance sheet or capital terms - if that promise is broken.

    ReplyDelete
    Replies
    1. Actually, I'm sort of in your boat here JKH, which was the motivation to my original inquiry. I also haven't aired any concerns about the CB's ability to defend price stability, but I share your concerns as well (I'm taking that as given for the sake of argument).

      I agree it feels more like an undertaking than what's typically perceived as a legal liability. But I think JP's point is that whatever it is, it's some type of promise or commitment relevant to the value of the fiat money, and what JP is interested in is identifying a promise that imparts value to the paper. If people believe the CB can establish and defend the value of a piece paper and will do it, then he argues that paper should assume value. The question of whether the CB can actually deliver on its promise is critical, as it factors into the value of the promise.

      Separately, your point about 'certainty of fulfillment' as a qualification for being a liability was something I was original thinking about. But I'm not sure how valid that is. Tons of things that are liabilities have uncertainty attached to their fulfillment (e.g., loans). So do you mean this in a different sense?

      Delete
    2. (In the previous thread, I was originally challenging him on very narrow balance-sheet terms, but he clarified that the kind of liability he's talking about is not really one in balance-sheet terms. It's more the pure nature of the promise CBs make to defend the value of currency. He also admitted that, according to his beliefs, if CB's abandoned a price stability target in favor of something like full employment, then the money could lose value. In other words, in his mind, all of this very much revolves around the promise to specifically defend the value of money.)

      Delete
    3. Hi ATR,

      In general, I'm in favor of keeping balance sheet analysis pristine from a financial analysis perspective, and using that in context when wider economic discussions are at issue.

      I intensely dislike some of the economic analysis I've seen that treats banknotes as "equity", and that sort of thing. And I gather some fairly important papers have been written on that. In my view it's all wrong as an approach that should instead be consistent in the integration of accounting and finance with economics.

      I think this is an excellent post, but I do see a mixture of at least two different categories of the notion of liability or IOU. The idea of maintaining a smoothly running operational machine that includes the exchange of banknotes for reserves and vice versa is quite different from the strategic objective of maintaining the currency's value.

      Delete
    4. "I think this is an excellent post, but I do see a mixture of at least two different categories of the notion of liability or IOU. The idea of maintaining a smoothly running operational machine that includes the exchange of banknotes for reserves and vice versa is quite different from the strategic objective of maintaining the currency's value."

      I agree here - they're different categories of liability/IOU. I think it would be helpful to make this clear, assuming JP Koning doesn't have any objections.

      Delete
    5. Hi JKH, your points about the criteria that make some promises a true balance sheet liability (and others not) are well taken.

      I was purposefully being vague in using the terms promise, IOU, and liability. The main viewpoint I'm rebutting (in both this post and the previous one) is the idea that bank notes require no undertaking on the part of a central bank whatsoever. Ever since Don Patinkin, the mainstream view has been that intrinsically useless, unbacked, and inconvertible notes could suddenly be propelled into circulation at a positive value and stay there. I won't go into the specifics of this view because George Selgin has done it so well here.

      The examples I've provided of IOUs/undertakings/liabilities would seem to provide counter evidence to the Patinkin view. Some sort of IOU must be provided in order get banknotes to circulate at a positive value and keep them there. Hopefully this point gets through despite the sloppiness of my categorization of IOUs/liabilities.

      Delete
  5. The best analogy I can think of for non-convertible bank notes is a stock like Berkshire, which is profitable but hasn't paid a dividend since the 1960s and isn't expected to pay a dividend for an indefinite period of time. Dividends are why stocks have value, so how can Berkshire stock have a positive value if it's non-dividend-paying? In this case, most people would have no trouble distinguishing between the ability to pay a dividend and whether dividends are currently being paid, or have historically been paid.

    Similarly, bank notes can be non-convertible for an indefinite period of time without ever losing a liability nature. The important thing is whether the bank could offer convertibility, not whether convertibility is currently being offered. Or, another way of looking at it, the important thing is whether the bank could reduce the quantity of money, not whether it ever has or is expected to.

    ReplyDelete
    Replies
    1. There are certainly some similarities there, although Berkshire stock is a claim to compounded profits over time whereas a banknote is nominally fixed. I like the analogy to perpetual senior bonds, where the interest return has been replaced by a liquidity return.

      Delete
  6. If I understand correctly, you are saying that central bank notes are given a floor to their value because they will at least be convertible to the assets held by the central bank.

    So circling back to your last post about bitcoin's bootstraps: isn't something like bitcoin given a floor value because it can at least be convertible to the goods which firms offer in exchange for bitcoin?

    The counterargument to this I anticipate is, "true, but the btc/usd exchange rate can suffer such that the goods/btc exchange rate can fall to zero. Therefore a bitcoin will lose purchasing power due to private firms asking for an unreasonably large quantity of bitcoins in exchange for the goods they offer."

    So the follow-up question I have is: Where in the documented legal framework that you are referring to here, does the central bank set an exchange rate of its notes to its assets? Is it that exchange rate fixed or floating? Otherwise the notes can suffer from exchanging for too few central bank assets, which is the exact same problem with bitcoins exchanging for too few private firm assets.

    ReplyDelete
    Replies
    1. "...isn't something like bitcoin given a floor value because it can at least be convertible to the goods which firms offer in exchange for bitcoin?"

      No firm has set a fixed quantity of goods into which they'll convert bitcoin. If bitcoin falls, they all raise their price. A central bank, on the other hand, sets a fixed quantity of hypothetical CPI baskets that banknotes are capable of purchasing. They do this indirectly by purchasing/selling government bonds. (Central banks usually don't encode inflation targets into their legal frameworks). As for the exchange rate between notes and assets (usually bonds), this floats in a way to ensure that the purchasing power of notes never falls by more than 3% fall each year.

      Delete
  7. Good thing it's pure unbacked fiat -- with only the most basic of currency-maintenance obligations. CB shareholder equity is an atavism as a consequence.

    Bonds -- the other asset side of the central bank sheet -- are the worst possible instruments for a CB to hold, if their shareholder equity matters.

    If a CB is successful in stimulating NGDP via base money production, then the mark-to-market value of the bonds declines. If the CB lowers NGDP, then shareholder value expands.

    There's a deflationary bias to CB's holding bonds worthy of international-gold-standard-immunization critiques.

    ReplyDelete
  8. Great post! It left me wondering what the believers in fiat money could possibly say to defend their positions, but Nick stepped up to the plate with this:

    "Nowadays, central banks promise to ensure that the value of notes doesn't fall more than x% against a CPI basket."

    1) That, to me, is the only sense in which Bank of Canada notes are a liability.

    2) And they are a negative real interest rate liability.

    3) Most of the assets of the BoC are themselves IOUs for banknotes. If the banknotes were worthless, so would be those assets."

    Reply to #1: I remember when Nick denied that BOC notes were a liability at all. But suppose that a central bank backed $100 of notes with assets that were worth 99 oz, and yet stood ready to buy back any note for 1 oz if the value of the notes started to fall against silver. There would be a run on the bank, with the first 99 notes being paid in full and the last one left worthless. In other words, any bank that tried to support the value of its notes but didn't FULLY back its notes would collapse.

    Reply to #2: Some entity could issue 10 year bonds on which it paid 5%, 1 year bonds on which it paid 2%, 30 day bonds on which it paid 0.5%, and zero-day bonds on which it paid -2%. In every case, customers are willing to accept lower yield for more liquidity. There is nothing exceptional about currency paying negative interest, and nothing that would make one question their status as a liability of their issuer. The currency would also not be net wealth to the nation.

    Reply to #3: Correct. And if GM held some of its assets in call options on GM stock (or anything else whose value depended on GM stock value), then you'd have a feedback problem, where GM loses some assets, so GM shares fall a little. But then the calls fall as a result, which makes the stock fall more, etc. The fact that there's feedback does not mean that GM's shares are not backed by GM's assets.

    ReplyDelete
    Replies
    1. Mike, thanks for jumping in. I too remember when Nick denied that BOC notes were a liability at all.

      Nick, would you say your views have changed a bit on this?

      Delete
    2. I think Nick agrees that assets are somewhat important since a central bank needs something to buy back its issuance. See his comment below, for instance:

      http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/10/liquidity-bubblesponzischain-letters-and-money.html

      But presumably if no more than 70% of a central bank's assets were destroyed (30% seems to be his magic number), it wouldn't matter too much to Nick... it only needs a few assets.

      In any case, I think we all agree that some sort of IOU exists. We can debate about the nature of that IOU till we turn blue.

      Delete
    3. JP you are probably correct regarding Nick's views: but I know for certain he did a whole post on why it's a mistake to talk about BoC notes as a liability of the BoC. Here it is:

      http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/03/is-modern-central-bank-money-a-liability.html

      (he's got a good search engine on his site)

      Delete
    4. Nick's views have changed, much to his credit. But I for one would like to see him, or anyone else, address my points # 1, 2, and 3 above.

      Delete
    5. Tom, something must have happened between March 2012 and October 2012. I think it was this post, one of my favorites:

      http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/04/from-gold-standard-to-cpi-standard.html

      Delete
  9. Do you think that fiat money where the monetary base is not a liability is possible, or do you know of such historical examples? The reason why I'm interested is that I argue in my recent paper (which I sent to you) that the presence or absence of a contractual relationship (e.g. a liability) is relevant for classification of money, but I omitted fiat money in my analysis because I didn't have a good explanation for it. If fiat money was always a liability, that would actually be better for my theory (well it's not really mine, but I think it's a good way of looking at money).

    ReplyDelete
    Replies
    1. Peter, I can't really think of any examples of paper currency that wasn't some sort of liability. There may have been "forced" currencies -- paper that circulated on pain of death. But this wouldn't strictly be fiat money since this sort of paper had underlying non-monetary value--it would have spared its user from execution.

      Delete
  10. In practice when a central bank has their reserves depleted by either confiscation by the government or attempting to support a peg that they can not defend, then the value of the currency goes down. You could watch this over the last year with Argentina and Venezuela. If it were just fiat money the reserves would not matter; however, the reserves clearly do matter to the value of the currency in these cases or any others I have looked at.

    ReplyDelete
    Replies
    1. Vincent:

      Do you have some good sources explaining the Argentina/Venezuela episodes?

      BTW: The bigger problem is loss of assets, not loss of reserves. A bank with adequate assets can always get more reserves, but a bank with inadequate assets will fail no matter how big a fraction of its assets is held as reserves.

      Delete
    2. I'm confused by the reserves comment too. Do you mean foreign cash reserves held as assets? The usual way to describe CB reserves is as liabilities of the CB. That's where the confusion comes in for me... foreign reserves would be assets though.

      Delete