Friday, March 1, 2013

Orphaned currency, the odd case of Somali shillings

A few weeks ago, David Beckworth egged me on to write about Somalian currency. I can't resist—it's a fascinating subject. The material I'm drawing on comes from Luther & White (2011), Luther (2012), Symes (2005), and Mubarak (2003)

Orphaned banknotes

When Somalia collapsed into civil war in January 1991, the doors of the Central Bank of Somalia were blown apart, its safes were blasted, and all cash and valuables were looted.* But something odd happened—Somali shilling banknotes continued to circulate among Somalians. To this day orphaned paper shillings are used in small transactions, despite the absence of any sort of central monetary authority.

The strange case of circulating Somali shillings forces us to ask some fundamental questions about money. If the Federal Reserve and all other branches of the US government were to be suddenly swallowed up into the sea, would Fed banknotes, like Somali shillings, continue to be used? What forces conspire to keep coloured squares of paper in circulation when their original issuer has long since expired?

According to Luther & White (2011), the shillings' continued circulation within the context of a collapsed state casts doubt on the universality of the chartal theory of money. According to chartal theory, the requirement that people pay taxes with government-issued bits of paper is what drives the positive value of these bits. Since a world with no state is a world with no taxes, the continued use of shillings means that something other than tax acceptance must be driving their positive value. [As an aside, I'd note that the ongoing circulation of bitcoin also contradicts the taxes-only theory of chartal money. After all, bitcoin isn't used to pay taxes, it's used to avoid taxes].

Luther & White give what seems to me to be a very Misesian explanation for the shilling's continued positive value: the "inertia of historical acceptance". According to Ludwig von Mises's regression theorem, outlined in Theory of Money and Credit (1912), people's expectations about the value of bits of paper may "regress" into the past. An intrinsically useless bit of paper is valued today because it had a positive value yesterday, and it had a positive value yesterday because it did the day before, all the way back to day 1 when that useless bit of paper was anchored to some commodity.

This process is described as a coordination game in Luther & White. Though the central bank had collapsed in January 1991, a Somalian trader might choose to still accept shillings the day after the collapse because he had accepted them the day before and he knew that others had accepted them too. In a self conscious manner, the trader would also know that other traders knew that he had accepted them. Expectations about expectations about expectations, a Keynesian beauty contest of sorts, was sufficient to drive the use of shillings beyond the day of the central bank's demise.

Counterfeit notes and contingent redemption

Let's add some more texture to our example. Not only did old shillings continue to circulate, but several new issues of counterfeit notes joined them. These counterfeit 1000 and 500 shilling banknotes were created by warlords and businessmen subsequent to the country's collapse. Although the counterfeit notes had the same design as the pre-1991 legacy notes, small differences allowed for differentiation. According to Luther (2012), of the five known forged issues, four have the date 1996 printed on them, long after the central bank ceased to exist. The counterfeits dated 1996 are signed by central bank governor Ali Abdi Amalow who, having been appointed in 1990, had never held office long enough to have his signature affixed to genuine Somali notes. Even without these imperfections, fake banknotes would have been instantly recognizable to anyone—they would have been crisp and clean relative to the limp and dirty legacy issue.

Despite being easily differentiable, Somalians willingly accepted counterfeit 1000 and 500 shilling notes. Not only did they accept them, they considered them to be fungible with real 1000s and 500s. In other words, the market refused to place a discount on the fakes.

Mubarak (2003) offers a few reasons for the acceptance of counterfeits. Any reticence the public may have had concerning the legitimacy of counterfeit note was overcome by A) coercion on the part of issuing warlord; B) a financial inducement to accept new notes including an initial 5% discount to the price of legacy notes, and C) the claim that new notes were liabilities of the Central Bank of Somalia. According to Mubarak the latter instilled a "general public belief that the forged... banknotes must eventually be upheld and honoured when effective national government comes to power."

This last detail is interesting because it might explain not only why counterfeits were accepted, but also why legacy Somali banknotes continued to circulate after the Somali state collapsed. Upon the eventual reconstitution of the Somalian state, a new central bank would most likely be created. This newly recapitalized Central Bank of Somalia would hold a stock of assets funded, say, by the IMF.  The central bank might take upon itself the original central bank's note liability No longer orphans, old banknotes could now be convertible into deposits held at the new central bank and, insofar as the central bank targeted inflation via open market operations, these deposits would in turn be convertible into genuine backing assets.

Thus the promise of redemption by a not-yet existing central bank may have been enough to give present shillings, both genuine and counterfeit, a positive value.

This means that any changes in the purchasing power of shillings might be due not only  to variations in the quantity of outstanding Somali media-of-exchange. Reassessments of the probability of a central bank both being created and honouring the previous note issue would also affect their purchasing power. For instance, should the Transitional Federal Government, a government in-waiting of sorts, succeed in consolidating its position in Somalia by winning a key battle, the odds of redeemability would increase, as would the value of shillings.

Historically orphaned currencies

This tension between fiat-paper-as-redeemabale-financial-asset and fiat-paper-as-medium-of-exchange is an old one. It was at the centre of one of the greatest monetary debates of the 19th and early 20th century, a dustup that involved luminaries like Irving Fisher, Knut Wicksell, Laurence Laughlin, Benjamin Anderson, Ralph Hawtrey, and Ludwig von Mises. The discussion centered on the irredeemability of US Greenbacks.**

Greenbacks, which had been issued in 1861 to pay for the Union Government's expenses, were initially 100% redeemable in specie. The redemption clause was suspended later that year and subsequent issues of greenbacks didn't even claim to be convertible into gold. Greenbacks quickly fell to a large discount relative to the metal. By August 1864 the discount had hit its widest point as the paper traded at 38 cents on the gold dollar.***

As with Somali shillings, economists were curious about these seemingly-orphaned liabilities. Why were greenbacks still accepted? What governed their value? Wicksell, Mises, and Hawtrey held that if there is a demand for the medium-of-exchange as such, this demand would be sufficient to give value to whatever instrument was established by custom as the medium-of-exchange. Thus the continued acceptance of greenbacks at positive values was mostly due their already-favorable marketability.

Countering this panel of illustrious economists was J. Laurence Laughlin. Though Laughlin doesn't attract the same brand recognition as do these other long-dead economists, in his day he was considered to be America's leading monetary economist. In his book The Principles of Money (1903), Laughlin outlined the view that greenbacks should be treated like non-dividend paying common stock. Just as a stock might have some positive value now due to potential dividends several years down the road, the continued positive valuation of greenbacks was due entirely to the possibility of their future redemption.

To illustrate his point, Laughlin drew attention to the market's reaction upon the success of the Union Goverment in the Battles of Gettysberg and Vicksburg. After these battles the greenback's discount to gold dramatically narrowed, presumably because these victories were perceived by the market as increasing the probability of future redemption of greenbacks in specie. Laughlin also mentions the passage of the Redemption Act of 1875, in which the government declared its intention to redeem all greenbacks come January 1, 1879. Though still trading at a discount to gold, greenbacks began to steadily appreciate towards par as this date approached, just like a t-bill approaching maturity. Thus Laughlin's redemption theory held up nicely to the evidence.

While Wicksell described Laughlin's theory as "perverse and fantastic," it was hard for Mises, Hawtrey, or Wicksell to deny that the expectation of future redemption didn't have at least some effect on greenback's value. In his book the Value of Money (1917), Benjamin Anderson struck a middle ground between all parties. Anderson held that Laughlin was right to treat greenbacks like any other asset whose value was dependent on eventual redeemability. But Anderson also thought that Mises and the rest were correct to put an emphasis on greenback's unique monetary function. Anderson combined these views by illustrating how greenback's "money-use" would be captured by the market as an extra bit of ascribed value on top of redemption value, or a liquidity premium.

Just to make things more complicated: New Somali Shillings

Back to the Somali shilling. The last feature of the Somalian monetary economy that I find interesting is the presence of yet another media of exchange — the "New Somali Shilling".**** Just prior to the January 1991 collapse, Somalia had been experiencing accelerating inflation. The Somali government drew up plans to replace existing shillings with the New Somali shillings. Each New Shilling was to be worth 100 old shillings, and notes were to be printed in 20 and 50 shilling denominations

 According to Symes (2006), the first shipment of New Shillings arrived in Somalia several months after the collapse of the state. Ali Mahdi Muhammed, a factional leader in Mogadishu, seized several billion of these official New Shillings and spent them into the economy in November 1991. While New Shillings did gain acceptance, their use was limited to a small area in North Mogadishu—the area controlled by Ali Mahdi Muhammed. Nor were they fungible with the legacy notes. 500 worth of New Shillings, for instance, was not worth 500 worth of old shillings. Nor did New Shillings circulate at the pre-1991 intended value of 100 old shillings to one New Shilling. At the time of Mubarak's article, a New shilling was worth only a third of an old shilling.

This challenges Luther & White's theory that Somalians accepted shillings because of their previous experience with them. Why did Somalians accept New Shillings at all if they were not accustomed to doing so? Why not refuse and continue to trade in legacy shillings? Legacy shillings circulated in Mogadishu at the time, so traders would have presumably had a choice. Why take a bet on an untested currency?

New Shillings also challenge the Laughlin-esque theory that redemption underpins the positive value of fiat paper. If Somalians universally valued shillings because of their future redeemability, why would they not place a higher value on New Shillings relative to old? After all, according to its stated plan, the Central Bank of Somalia was willing to convert 100 old shillings into 1 New Shilling. If it was believed that a newly chartered central bank would take on the liability of Somalia's counterfeit notes, wouldn't that same central bank uphold a commitment to value New Shillings at a far higher rate than old ones? Why then do New Shillings trade a third the price of old shillings rather than at a multiple of 100?

As much as I dislike to leave more questions than answers, that's about all I can do for the time being. One major data point is yet to come. What actually happens to legacy and counterfeit notes when a new central bank begins to operate? The Central Bank of Somalia's website says this:
Towards fully assuming its functional and institutional responsibilities, the Bank has completed the reconstruction of its Headquarters in Mogadishu while the branch in Baidoa has been established and is already functioning with personnel in place.
Perhaps we don't have long to wait.

*Abduraham in Luther (2012)
**Debate also surrounded the irredeemability of Austrian gulder banknotes. The Revolution of May 1848 resulted in the withdrawal of the convertibility of Austrian gulden banknotes into silver. The discount on gulder notes relative to silver averaged around 15% for the first 10 years and then dipped to an average of 28% from 1859 to 1865. Laughlin pointed out that by the 1860s speculation began to grow that Austria would switch from the silver standard to a gold standard. If the banknotes, still irredeemable, were to be honoured, it became increasingly evident that redemption would be in gold, not silver. At the same time, large silver discoveries were driving down the value of silver relative to gold. As a result of these forces, the market value of notes rose back to par with silver. As the market became progressively more confident that the notes would be redeemed, and redemption would be in much more valuable gold, gulder notes eventually exceeded their stated silver value. Laughlin's theory seemed to be borne out by the facts--the value of orphaned notes was dictated by their future potential redeemability.
*** From A History of the Greenbacks (1903), by Wesley Clair Mitchell
**** Mubarak refers to this instrument as the Na' Shilling.


  1. "While New Shillings did gain acceptance, their use was limited to a small area in North Mogadishu—the area controlled by Ali Mahdi Muhammed."

    That would be consistent with the view that the warlord gave them value because he accepted them in payment of protection money (i.e., taxes). The tax-backing theory still looks OK. (Subtle difference between the tax backing view and chartalism: Chartalism says taxes create a demand for money, while the tax backing view says taxes are an asset of the government, and as such serve as backing for the money, along with the government's other assets.)

    Also, I remember reading that the value of Somali shillings was close to being equal to the cost of printing the paper.

    1. I agree that the acceptance of the New Shilling would be consistent with both tax-backing and chartal theories, however, there's no data to indicate that Ali Mahdi Muhammed actually took on the New Shillings as a liability, or that he required them to be used to discharge obligations to his faction.

      It could also be indicative of the money-as-a social-contrivance theory. A tribe might be small and close knit enough to decide to give value to some intrinsically useless bit of paper.

      "Also, I remember reading that the value of Somali shillings was close to being equal to the cost of printing the paper."

      Yes, that's in Luther 2012. Very good paper. Luther calls it an "ink and paper" standard.

  2. Thanks for doing the post JP. So what are thoughts about the shillings become effectively a commodity money, "ink and paper" standard?

    1. Hi David.

      On the supply side, Somalia certainly seems to have a commodity standard in which production is governed by cost & competition. I like Luther's account of this in his 2012 paper.

      It's a little bit weird on the demand side. In a cattle standard, for instance, the demand for a marginal head of cattle could be derived from consumptive use or monetary use. Same with a gold standard, in which demand for jewelery intertwines with demand for coins. In Somalia, I don't think there is any consumptive demand for shillings i.e. there is no demand on the margin for shillings as writing paper, or fuel for fires... there is only monetary demand.

  3. Somali shillings circulate because they already are the most liquid good in Somalia, and the state hasn't replaced them with new money. This is a rare case, where Rothbard and Krugman agree.

    If the new central bank issues new money and sets a fixed exchange ratio with the old schillings, the old schillings will be removed from circulation (Gresham's Law). With the New Somali Schillings, it looks like the process got stuck in the middle.

    I would say that there are two different questions:
    - why do the Schillings act as money?
    - why is the exchange ratio the way it is?

    These need to be treated separately.

    The reason why counterfeits are treated as fungible becomes apparent if you spend some time with the concept of immaterial goods and copying. Whether the copy is or isn't treated as a perfect substitute depends on the appraisal of the user, not the issuer. Typically, when the state (issuer) finds that you have a counterfeit note, it is confiscated and you might be even charged with counterfeiting. That motivates people to distinguish between them. In Somalia, this restriction is absent. It looks like this has a significant effect on the substitutability of the notes (this is just a remark and is not supposed to be interpreted as a general rule).

    1. "Somali shillings circulate because they already are the most liquid good in Somalia"

      As I pointed out in the post, I don't think this works for the New Shilling. New Shillings weren't liquid at all, yet they went into circulation. Why would the market have allowed this to happen if old shillings were already one of the more liquid instruments in Somalia? Something else is at force here. Unfortunately we don't have enough data on New Shillings.

    2. I would need to read up the details (I read the old Luther article but not the new one), but my guess is that the replacement process got stuck in the middle, before the old shillings were fully displaced. There have been plenty of cases where the state replaced an old currency with a new one, so we have a idea about how it works if the migration is successful. Maybe the case of New Somali Shilling is an example of an unsuccessful migration. That could explain the floating exchange rate and limited geographical acceptance.

    3. If you want the Mubarak paper, email me. The rest are available on the web.

      I've got your thesis open on my computer and will get to it soon!

  4. One factor you didn't mention at all is the economy's need for a medium of exchange. You don't need a hard theory of why a note is perceived to be valuable if there is a scarcity of alternatives!