Sunday, March 23, 2014

Dismantling a central bank

In a previous post, I made the point that banknotes aren't Samuelsonian bubble assets, say like a chain letter or a ponzi scheme or bitcoin. Upon the dismantling of a central bank, each note has a senior claim on a central bank's remaining assets. Rather than being mere "oblongs of paper", as Samuelson described them, banknotes occupy the very top of the capital structure hierarchy, above stock and bonds. This quality of being "well backed" might be sufficient for notes to trade at a positive value in the first place, and also help explain subsequent fluctuations in the purchasing power of those notes. In this post I revisit these ideas.

The process of dismantling a central bank would go like this. Imagine that the Reserve Bank of Fiji announces that it will wind up operations next week and recall all notes. Its assets consist entirely of Fijian dollar-denominated bonds. With the eminent end of Fijian dollars, the entire Fijian economy will have to quickly re-denominate existing debts and contracts into a new unit of account, say U.S. dollars. After this redenomination, the Reserve Bank of Fiji now finds itself holding a large quantity of U.S. dollar-denominated debt. It proceeds to sell these bonds, as well as its printing presses and premises, for dollars, and then uses dollars to cancel all Fijian notes. The Reserve Bank of Fiji is no more, and neither are Fijian dollars.

Because the central bank's hypothetical dissolution would result in noteholders ending up with some ultimate quantity of U.S. dollars in their pockets, Fijians can use this terminal value as a basis for computing the present value of Fijian banknotes. They would go about doing this computation in the same that they would with a stock and bond, both of which also have hypothetical terminal values, or liquidation values.

The inestimable Mike Freimuth, who blogs here, pointed out that there is a problem with this. How can Fijian noteholders actually quantify the amount of dollars to which they are entitled upon dissolution? This is easy if the Reserve Bank of Fiji is already on a dollar standard (say it redeems notes for US$1). Once all of the central bank's assets have been sold for U.S. dollars, each noteholder would get US$1 until all Fijian noteholders had been satisfied, upon which less senior stakeholders like bondholders and shareholders would get their portion of the central bank's remaining stash of U.S. dollars.

However, if Fiji is on a floating standard, the task of valuing noteholders' claims gets quite thorny. While fiat notes may have a senior claim on assets upon dissolution, it isn't evident how many actual dollars this entitles noteholders too. And if their quota of remaining assets, or terminal value, is not known at the outset, how can Fijians arrive at a value for notes in the first place?

One answer to Mike's criticism is that Fijian note owners come up with their best estimate of how much dollars the central bank (or a judge) would decide to award them upon a hypothetical dissolution. If the market's collective guess is that no assets would be forthcoming, then Fijian dollars would be worth nothing upon windup. If Fijians on average assume that the sale of bank assets would net $100m, and they think a judge would award them half of that, then each Fijian banknote will earn a prorated share of $50m. If they think that the all assets will be awarded to them, then they'll get a prorated share of $100m.

While these expectations about the terminal value of notes would probably be sufficient to jumpstart the positive value of Fijian dollars in the first place, this isn't a very satisfying explanation for subsequent variations in the price level. After all, the Fijian price level on any given day would be a function of Fijians' many and divergent expectations concerning their as-yet-unstated share of some future pie. As noteholders take potshots at guessing what this unknown size of this slice would be, the Fijian dollar would fluctuate wildly, sort of like a penny stock. Penny stocks owners have wildly fluctuating estimates concerning their ultimate, and unfixed, share of the corporate pie, which is only determined after debtors have had their pickings.

However, the behavior of currencies is not like penny stocks—the former tend to vary only a little in price from day-to-day. To explain price level fluctuations, it would seem that something other than the terminal value of the Reserve Bank of Fiji's assets must be at work.

Here I would like to reiterate my point from an earlier post that once a banknote has been jumpstarted into having a positive value, the dissolution-value of a central bank assets will only have a distant influence on those note's subsequent value. By far the more immediate effect that central bank assets have on the value of notes emerges via their ability to be mobilized in the maintenance of price stability. By selling assets for notes and retiring them (or threatening to do so), a central bank can prevent the value of their notes from flagging. Alternatively, the income earned by those assets provides the central bank with the resources to pay more interest (on reserves at least), a feature that will also ensure price stability.

So it seems to me that any impairment to a central bank's assets will affect the value of notes not so much because it lowers their value come future dissolution, but because it limits the central bank's ability to repurchase notes and pay interest in the present so as to maintain their price target. If a central bank didn't provide a target supported by a repurchase facility, then the value of notes would be dictated by something like their terminal value—and prices would be much more volatile then they are now.


  1. Hi JP:

    "By selling assets for notes and retiring them (or threatening to do so), a central bank can prevent the value of their notes from flagging."

    That's true as long as the central bank has positive net worth. But suppose the Fijian central bank has issued 100 Fiji dollars, and its assets are worth only $90 US. The market will value 1 Fiji dollar=$.90 US. If the Fiji central bank then tries to support the Fiji dollar by paying $10 US to buy back 10 Fiji dollars, then the Fiji central bank will have $80 US backing 90 Fiji dollars, and the Fiji dollar will fall to $.89 US.

    Conclusion: When net worth is positive, the value of the Fiji dollar is set by the rate of convertibility maintained by the central bank. When net worth is zero or negative, the value of the Fiji dollar is determined strictly by backing.

    1. "The market will value 1 Fiji dollar=$.90 US."


      We know that each of the 100 Fiji dollars issued provides a claim of X on the US$90, but we don't the size of X. For instance, the market could value 1 Fiji dollar at US$.90, as you point out, but it could just as easily value 1 Fiji dollar = US$0.50, and that the remaining $40 goes to residual claimants. Because the size of the claim isn't specified, any price of 1 Fiji dollar between $0 and $0.90 makes sense.

  2. "...... the entire Fijian economy will have to quickly re-denominate existing debts and contracts into a new unit of account, say U.S. dollars."

    It seems to me like there would be two ways of doing this:

    1. An arbitrary decree is issued stating "One Fiji Dollar is now worth One US Dollar". Done.

    2. Real American dollars are obtained for Fiji assets so that a convertibility base is established. This requires a real exchange with holders of US dollars.

    Convertibility between currencies is usually done on the margin, with only small amounts of currency traded. It seems to me that converting a very large amount of currency, here the entire economy, would be a much more difficult undertaking.

    A version of this is apparently happening in Crimea. Winterspeak has an interesting post on this topic found at Crimea will need to convert from Euro to Ruble. I must note that Crimea does not have a central bank.

    1. The Russians will figure out a way to make it work, they just went through the process of ruble-izing South Ossettia and Abkhazia. And of course the Europeans went through the whole process in 1999. The decree, however, shouldn't be an "arbitrary" one, because if the rate that is declared overvalues one of the two currencies, then either the creditors or debtors will be out of pocket.

      A redenomination could also occur spontaneously. If the public is made aware that the national unit is about to go extinct, merchants will quickly evolve their own pricing standard. Either the market will create a new unit, or an existing unit will be selected, in the same way that Zimbabweans converged on $US in 2008 without needing a higher authority to tell them to do so.

  3. I totally agree with you that price stability is the key basis for the value of currency. If I put one years wages as dollars, I will still be able to get one years wages worth of stuff with them next week -That isn't so true with bitcoin or gold. I guess that assurance of price stability depends somewhat on other parts of the state beyond the central bank. Wage and supply contracts and taxes denominated in the currency also contribute and of course the legal system to enforce it.
    In reality when a currency system is wound down is it not more typical for money holders to be dispossessed? Am I right in understanding that after WWII, the Nazi era German monetary system was effectively written down with German people simply having to start over.
    So money has its value for the stability it provides when all is going smoothly despite the possibility that it may be declared worthless as a decree of a future political power (as happened in Germany when the Allies rebooted the post WWII German money).
    "On Sunday, June 20, the counters which had previously issued ration stamps began to hand out the new currency. Every German received 40 Deutschmarks that day, followed by a further 20 Deutschmarks one month later. That Sunday, all German citizens were equally rich or poor."

    1. Interesting data point.

      I think the following article gives a better account of the death of reichsmarks and debut of deutsche marks.

      "On June 18, 1948, the American and British occupation authorities announced that a new unit of currency, the Deutschmark (DM), would replace the Reichsmark (RM) on June 20, and would be the sole legal tender beginning June 21. Each German was given an allowance of 60 DM (Kopfgeld), exchangeable at a 1:1 ratio with RM, and all bank deposits exceeding 60 RM to be converted at a rate of 10:1. Half of these bank deposits were temporarily frozen, and employers were given an allowance of 60 DM per employee to cover the first week’s payroll."

      So reichsmarks weren't declared worthless, nor were people dispossessed. The redemption rate between the old and new mark was reduced after the initial 60RMs, but there was still a possibility of converting at a lower rate.

    2. Wikipedia also backs up the version you found. However having 90% of your savings wiped out still is quite an impact even if it is not 100%.
      Although 90% is not 100% it nevertheless goes along with the general point that you make that money is valuable for what it does now rather than because it is some infallible claim on those rare occasions when a monetary system gets wound down.