Tuesday, November 21, 2017

Zimbabwe and hyperbitcoinization

Art by Daniel Krawisz

The narrative that drives any speculative market needs constant fuel in order to attract new buyers. Bitcoin is no exception, which is why recent events in Zimbabwe have been recruited—albeit sloppily—by the bitcoin press to provide more fuel.

The facts of the matter are this: if you head over to Golix, Zimbabwe's only bitcoin exchange, you'll see that bitcoin last traded at $13,800 whereas its price on an American exchange like GDAX is $8260. That's a difference of around $5000. Strange, right?

The bitcoin press, and I'll pick on Zerohedge here, has interpreted this divergence to mean that bitcoin usage is skyrocketing in Zimbabwe. Zimbabweans are seemingly so desperate to get their hands on some bitcoin that they are willing to pay $5000 more per coin than they would pay if they bought on an international exchange like GDAX.   

Recall that one of bitcoin's earliest and most potent use cases was to provide unbanked Africans with an efficient way to transact. The effort to bring bitcoin to Africa has significantly underperformed the earlier hype—but nevertheless the dream of helping out the poorest continent still beckons. Well, we've finally got a real-world case of bitcoin being used by Africans. And if Zimbabweans have adopted the stuff as money, goes the story, then it's only a matter of time before other developing countries and then the whole world goes full hyperbitcoinization. This last term refers to an oft-quoted idea originating from Daniel Krawisz that currencies will fail to compete with bitcoin, leading to a rapid bitcoinization of the world. Think of it as dollarization, except really fast.


The truth is that Zimbabweans are not paying $5000 more per bitcoin than everyone else. As I've written many times on this blog, and recently at BullionStar (see my postscript at the bottom of this post), Zimbabwe is no longer on a U.S. dollar standard, having had a new Zimbabwean currency thrust on it by the government over the last year or two. This new currency was supposed to be fully convertible into U.S. dollars, a promise that has proven to be illusory. It has since slipped to a large discount to the dollar.

So when Zimbabweans buy bitcoin for $13800—they aren't paying with U.S. dollars, they are paying with this new unit. The Zimbabwean price for bitcoin therefore deviates from the U.S. dollar price  for the same reason that the Mexican bitcoin price of Mex$153,000 deviates from the U.S. dollar price—Zimbabwe, like Mexico, has its own freely-floating currency. This explanation for Zimbabwe's peculiar bitcoin price certainly makes for less interesting headlines than the explanation put forward by the bitcoin press.


Here's why I find the price of bitcoin in Zimbabwe interesting. The prices of easily transportable items with high value-to-weight ratios should trade at the same real price all around the world. Included in this category are things like human blood, diamonds, plutonium, gold and silver, heroin, $100 bills, stock certificates, and of course bitcoin.

From Evil Mad Scientist, the Monetary Density of Things

The law of one price prevails because if the price of gold in a country like Zimbabwe falls below the world price, than arbitrageurs will buy gold in Zimbabwe and sell it overseas until the price has converged back up to the world price. If the gold price rises above the world price, they'll buy overseas and export it to Zimbabwe until the price falls.

Items with high-value-to-weight ratios like gold and bitcoin provide an opportunity to infer currency exchange rates. Because a bitcoin trades for Mex$153,000 in Mexico and $8260 in the U.S., we don't even have to visit a foreign exchange website to know that the exchange rate is about 19 Mexican pesos to US$1. Just divide $153,000 by $8260. Likewise in Zimbabwe. Given a bitcoin price of $13,800 in Zimbabwe and $8260 in the U.S., we can safely assume that the exchange rate is around 1.70 Zimbabwean currency units to US$1.

This ability to back out an exchange rate using items with high value-to-weight ratios is especially handy for researchers and reporters who are observing countries from afar that lack official venues for trading currency. In Zimbabwe, the market exchange rate is set unofficially, on street corners and such. Without an on-the-ground data gathering network to canvas street corners, an outside observer can get a decent proxy for the exchange rate by gathering bitcoin prices on the internet.

In fact, inflation researchers like Steve Hanke have long-since been gathering data on high value-to-weight items to back out exchange rates, although they have typically used the prices of inter-listed stock rather than bitcoin for this purpose. In this post I described the inter-listed stock technique being used in Zimbabwe, and here I describe how it could be used in Greece and Ecuador. In the future, bitcoin exchanges may offer yet another way to formally gather unofficial exchange rates.


 Zimbabwe's autocratic leader Robert Mugabe was put under house arrest by the army on November 14. Let's use what we now know about bitcoin prices to observe what has happened to the exchange rate since then. The price of bitcoin on Golix has been stuck in a range between $12,000 to $15,000 even as the U.S. dollar price on bitcoin exchange GDAX has jumped by ~$1500, or 26%. Golix prices are incredibly variable, far more so than international prices, so I don't want to overstate my case, but it seems likely that the purchasing power of Zimbabwean currency—inferred from bitcoin prices—has actually improved since the coup. 

As I said in my BullionStar article (again, see note at bottom), this firming up of the Zimbabwean currency may indicate that markets see some improvement in the odds that a stable new regime emerges in Zimbabwe, one that enjoys acceptance by the international community. Due to its pariah status, the Mugabe government has been unable to get aid from the World Bank or IMF for many years now. If its status changes, a new loan could allow the country to a re-peg the Zimbabwean currency at a 1:1 rate with U.S. dollars.

The improvement in the exchange rate that I've inferred from bitcoin prices is corroborated by looking at the prices of other items with high value-to-weight ratios. As I pointed out earlier, Steve Hanke uses inter-listed stocks to back out exchange rates, in Zimbabwe's case the shares of Old Mutual which are traded both in Harare and London. The ratio between the two listings is known as the Old Mutual Implied Rate, or OMIR. Gareth from Twitter, who knows a lot about Zimbabwe's financial situation, sends me the following chart of OMIR (the green line).

You can see that there has been a big decline (35.1%) in the Zimbabwean price of Old Mutual (white line), but almost no change in its London price (orange line). As a result, the green line—the ratio of the London price to the Zimbabwe price—has moved higher. Which means that the purchasing power of Zimbabwean currency has improved since Mugabe's arrest, albeit just by a bit, corroborating what we already learned from the Golix-to-GDAX bitcoin ratio.


None of this is evidence of hyperbitcoinization in Zimbabwe. In fact, if you spend some time on Golix you'll quickly notice that volumes are very thin and the trading range incredibly wide. Over the course of a few hours, I saw the price on Golix rise and fall by $1,500 whereas the U.S. price on GDAX has been relatively immobile. Bitcoin is attracting some traders, but it hardly seems popular. Any Zimbabwean who wants a safe haven has probably been buying U.S. dollars on the street corners. Sorry folks, but as far as I can tell Africa still lacks a bitcoin use case.

P.S. I am happy to announce that for the first time I'm being paid to blog. The folks at BullionStar will be hosting blog posts from me over the next few months. Expect the posts to be on some of the same topics I blog about here; monetary policy, gold, bitcoin, and more. My first effort gives a quick overview on the Zimbabwean monetary situation. I am always looking for more opportunities for paid blogging in my general area of expertise; contact me if you have any leads.

Monday, November 6, 2017

The bootstrapping of Thorne, Magic Money, and Cyberbucks: three pre-Bitcoin monetary experiments

Bitcoin boasts many technical achievements, but none is more interesting to me than they way it was successfully bootstrapped. How did a small group of cypherpunks—activists interested in widespread use of cryptography and digital currency—manage to get an intrinsically valueless token to have a consistently positive price? Hal Finney, a cryptographer and early adopter of bitcoin, put it this way in 2009:
"One immediate problem with any new currency is how to value it. Even ignoring the practical problem that virtually no one will accept it at first, there is still a difficulty in coming up with a reasonable argument in favor of a particular non-zero value for the coins."
The bootstrapping of bitcoin seems to have been achieved with some care. William Luther has gone through old bitcoin message boards to show how early adopters, including Finney and bitcon-creator Satoshi Nakamoto, coordinated to 'enter the network' at the same time, thus generating a positive value for worthless bitcoin tokens. A token that is already valuable, perhaps because it is useful for some non-monetary use like jewellery, or because it is directly convertible into an already-existing money, is much easier to launch than one that isn't already valuable. Bitcoin didn't have the benefit of non-monetary usefulness.


By way of Timothy May's Cyphernomicon, I recently discovered that bitcoin wasn't the first attempt by cryptographers to launch an intrinsically worthless digital token into positive-value space. Similar bootstrapping attempts occurred back in a previous era of digital currency experimentation, the mid-1990s.

In 1993 the extropians—a group that believes in the technological possibility of immortality (among other things)—set up an experimental market called the Hawthorne Exchange where individuals could trade units of reputation. There seems to be some crossover between extropians and cypherpunks with the reputations of folks like Timothy May and Nick Szabo, both key contributors to the Cypherpunks electronic mailing list, being listed on the Hawthorne Exchange. Trades were made using the exchange's own native currency called thorne, which had a fixed supply. Not only did the extropians succeed in generating a positive price for twenty or thirty reputation tokens, but by extension the native currency—thornes—was also bootstrapped.

As part of the experiment, people began to sell stuff for thornes, including copies of digital cash papers and old books. They made bets in thornes and even established a U.S. dollar price for the nascent digital currency (it was somewhere between 100 and 1000 thornes per dollar).

The problem with the whole endeavour is that—as Hal Finney would point out not long after it had begun—the tokens were essentially worthless. By convention each unit was supposed to represent a person's reputation, but there was no independent force that could possibly make a token correspond to a reputation:
"It is important to understand that Thornes are not like dollars. Unless HeX shares can be given a grounding other than the whim of their owners, the market will surely collapse, because there is nothing to support it."  
Finney would be proven right, since the Hawthorne exchange was shut down sometime in 1994.


In their next effort the cypherpunks would bootstrap a set of play currencies that had been created using a toolkit called Magic Money, a digital cash system programmed by the pseudonymous Pr0duct Cypher and made available in February 1994. Here is Pr0duct Cypher in the introduction to the software:
"Now, if you're still awake, comes the fun part: how do you introduce real value into your digicash system? How, for that matter, do you even get people to play with it?

What makes gold valuable? It has some useful properties: it is a good conductor, is resistant to corrosion and chemicals, etc. But those have only recently become important. Why has gold been valuable for thousands of years? It's pretty, it's shiny, and most importantly, it is scarce.

Digicash is pretty and shiny. People have been talking about it for years, but few have actually used it. You can make your cash more interesting by giving your server a provocative name. Running it through a remailer could give it an 'underground' feel, which would attract people.

Your digicash should be scarce. Don't give it away in large quantities. Get some people to play with your server, passing coins back and forth. Have a contest - the first person who (breaks this code, answers this question, etc.) wins some digital money. Once people start getting interested, your digital money will be in demand. Make sure demand always exceeds supply."
From the cypherpunks mailing list we learn that over the course of the next few months four or five unique tokens were created using Magic Money, including Tacky Tokens, GhostMarks, DigiFrancs, and NexusBucks. As in the earlier case of thornes, an attempt was made to sell goods and services in these new currencies. One poster on the Cypherpunk message board offered to pay coders to write software with NexusBucks, and another tried to sell GIF art of tacky tokens.

After a flurry of activity, however, interest died off. "It appears that the Magic Money/Tacky Token experiment is not succeeding in producing an informal digital currency," wrote Hal Finney in May 1994. "People have offered services in exchange for this money but have had no takers." In a post entitled Why Digital Cash is Not Being Used, Tim May blamed the failure of Magic Money on the lack of items to buy with tokens and confusion about how to get them and send them. It's worth a read.


No sooner had the Magic Money experiment died when a new new opportunity for bootstrapping digital tokens emerged. David Chaum, an early advocate of privacy, had established a company called Digicash in 1989 to commercialize the use of blind signature technology in electronic currency. In a trial that was first announced in July 1994, Digicash offered the first 10,000 applicants one hundred free cyberbucks, or e$, up to a maximum issue of one million cyberbucks.

Despite the fact that these tokens were intrinsically worthless—they had neither commodity value nor could they be redeemed into U.S. dollars—people soon began to transact with them. On its website, DigiCash listed around 100 shops that accepted cyberbucks, including those that sold postcards and various types of information services. Zooko Wilcox-Hearn, who recently founded the anonymous cryptocurrency Zcash, offered to sell his PGP software for cyberbucks. A coding contest by the omnipresent Hal Finney offered cyberbucks as a prize and Adam Back, a cryptographer who is currently involved in administrating bitcoin, sold "export-prohibited" cryptographic t-shirts for a price of e$250. In the same way that a pizza was the first good to be bought with bitcoin, Back's t-shirts may have been the first to be bought with cyberbucks:

To Digicash's surprise, several primitive financial markets emerged to trade cyberbucks for genuine currency. On the Ecash Exchange Market, which was hosted on the website of company called Firecloud Solutions, a price of around five cents per e$1 was established (see image below), effectively valuing the entire market capitalization of cyberbucks at $50,000.

Source: A Common Currency System for Spontaneous Transactions on Public Networks

For those with long memories, the above Ecash market looks very similar to New Liberty Standard's bitcoin-to-paypal market, the first bitcoin exchange that was established in 2009.

The cyberbuck trial did not last. While there was plenty of discussion about the topic in 1995, there are only a few mentions of cyberbucks on the Cypherpunk mailing list in 1996, but almost nothing in 1997. When I asked Zooko if he still had cyberbucks, he told me he had long since lost his. Who knows? They might still be worth a lot as collector's items.

Like cyberbucks and the other mid-90s experiments, bitcoin began as a mere play thing among a small coalition of technologists interested in privacy. Why did bitcoin get successfully boostrapped while the others failed? How did one form of monopoly money spread over the entire globe while the others were never used by anyone other than a small band of cypherpunks?

One answer is luck. Perhaps nothing more than a fortuitous flap of a butterfly's wings in Brazil set the whole thing off. Another is experience. After three failed efforts to bootstrap electronic tokens, perhaps the cypherpunk community had developed a better understanding of what not to do to get the ball rolling. Hal Finney for one participated in all four digital currency experiments.

The technology was different as well. Because it utilized David Chaum's patented blind signature protocol, Magic Money was technically illegal, and thus unlikely to spread to more timid adopters. As for cyberbucks, once the trial was over the server running the software would have to be shut off, at which point there would be no way to verify cyberbuck transactions. Knowledge of this imminent shutdown would have handicapped the ability of cyberbucks to propagate beyond the core group of hobbyists. Bitcoin, on the other hand, used a decentralized (and unpatented) method of verifying transactions, so the threat of winding up the system was less salient.

I'm not sure these technical factors were as important as the different macroeconomic environments in which the various digital currencies were issued. Cyberbucks, Magic Money, and thorne all appeared when the global economy was humming along and interest rates were high. Owning these zero-yielding tokens meant that users had to make a large sacrifice. In 2009, interest rates around the world had fallen to near zero, so holding a digital currency like bitcoin did not involve forgoing much in the way of interest income.

If usage of an intrinsically worthless token is to spread beyond an inner clique of hobbyists, a whole army of dreamers and speculators has to be encouraged to jump onto the bandwagon. What better pool to recruit from than the ranks of unemployed and underemployed in the wake of the 2008 financial crisis? This pool of downtrodden simply didn't exist in the humming 1990s. Folks back then had no need for a bubble asset to get them ahead—they enjoyed full-time jobs and plenty of opportunity.

Perhaps we were all a bit innocent in the 1990s and didn't understand how much our privacy could be invaded by governments and corporations. Magic Money and cyberbucks, which promised  protection from these threats, arrived too early. When bitcoin was finally introduced, it may be that we had all become a bit wiser and thus more willing to endure the hassles of switching some of our wealth into cludgy digital currency.

Lastly, people weren't upset with the finance establishment back when thornes, Magic Money, and cyberbucks were being introduced to the world. While recessions had hit in the early 1980s and 90s, they weren't accompanied with large-scale financial meltdowns. But in 2009, the credit crisis and bailouts were just in the rear-view mirror. Many were furious with banksters, and justifiably so. Turning to bitcoin was a protest vote.

Friday, October 27, 2017

The evolution of the Federal Reserve's promises as recorded on their banknotes

Since they began to be produced in 1914, Federal Reserve notes have always had a promise or obligation printed on their face. But over time this promise has changed. I thought it would be fun to go through the evolution of the promise as an exercise in understanding how the U.S.'s monetary plumbing has changed.

The original 1914 series of Federal Reserve notes had the above stipulation printed on it. It's tough to read, so I've reproduced it in full below:
"This note is receivable by all national and member banks and Federal Reserve Banks and for all taxes, customs and other public dues. It is redeemable in gold on demand at the Treasury Department of the United States in the city of Washington, District of Columbia or in gold or lawful money at any Federal Reserve Bank."
There were really four promises here. The first was that all banks who were members of the Federal Reserve system would accept notes at their counter, as would the Reserve banks themselves—i.e. the twelve district banks. The second promise was that the government would accept the notes in payment of taxes. This is what I described in last week's post as twintopt, or the MMT/Wicksteed idea of tax receivability. The third promise was to uphold the gold standard, both the Fed and the Treasury being obliged to redeem notes with the yellow metal. And the fourth and final promise was to redeem notes with something called lawful money.


Let's focus on this last promise, which is the most curious. What does lawful money mean? The Fed gives a brief description in their FAQ, the summary of which is: it's complicated.

Before the Fed's founding in 1913, the U.S. issued a smorgasbord of different government currencies, as the chart below illustrates. The most important of these issues was United States notes, otherwise known as greenbacks. The Treasury was first authorized to issue them when the Civil War broke out, and only in 1994 was it relieved of its its obligation to redeem old United States notes with new ones.

The 1862 act that authorized United States notes declared them to be lawful money and a legal tender, phraseology that implied that to be lawful was distinct from being legal tender. While the act never explicitly defined lawful money, legal tender was already a well-known term—meaning any instrument that by default can be used to discharge a debt. So if James owns Judith $20, legal tender is any instrument that Judith cannot refuse to accept when James wants to discharge his debt.

People often assume that government money and legal tender are synonymous, but in actuality there are many examples of government banknotes that have not been legal tender. Take Scotland, for instance, where neither Scottish banknotes nor Bank of England notes currently have legal tender status. U.S. silver and gold certificates, two forms of circulating paper money that were issued by the Treasury through the 19th and 20th century, were not legal tender—at least not till 1919 in the case of gold, and 1934 for silver. Nor were National bank notes a form of legal tender. These were private banknotes issued by banks chartered under the National Bank Act (see chart above). The Treasury promised to accept National bank notes at par in discharge of most Federal taxes, effectively adopting them as their own IOU, yet refused to grant them legal tender status.

While there is no formal definition of lawful money in the 1862 act, we know that it wasn't necessarily legal tender, and we do know that greenbacks fell into said category. Thus a narrow reading of the lawful money promise on a 1914 Federal Reserve note simply meant that they were convertible into United States notes, which by chance happened to be legal tender.

A much broader definition of lawful money would emerge later. By 1935, Fed Chair Mariner Eccles included not only United States notes—greenbacks—but also silver certificates and National bank notes in the category of lawful money (see pdf). William McChesney Martin, who served as Chair from 1951 to 1970, defined lawful money as "any medium of exchange which frequently circulates from hand to hand as money under sanction of the law." (pdf) By this generous definition, lawful money referred to the entire mongrel collection of government currencies, including legal tender United States notes and non-legal tenders such as silver/gold certificates and National bank notes, and finally legacy notes no longer being printed including Treasury notes of 1890, fractional currency, and old demand notes.

The upshot is that the original promise to redeem Federal Reserve notes with lawful money effectively linked what was then a novel currency to the medley of recognizable government currencies already in use. This promise would have provided the public with much-needed continuity between the familiar and not-so-familiar. After all, the Fed was a new institution that had not yet earned credibility. Tying its obligations closely together with greenbacks and/or every bit of paper then in existence would have helped its cause.


Having dealt with the idea of lawful money. Let's go on to look at how the promises on Fed notes changed after 1914. I've included the 1928, 1934, and 1963 series:

You can see that the 1928 series no longer included either the first (receivability at banks) or second promises (tax receivability). This was probably for brevity's sake. Even though these promises no longer appeared on Fed notes, they continue to be enshrined in Section 16.1 of the Federal Reserve Act. So when you do your taxes in 2018, the Treasury is obligated to accept notes as payment, just as they were in 1914, even though this isn't indicated on the face of a banknote.

When the 1934 series was printed, the third promise—to redeem in gold on demand—was dropped, both from the face of banknotes and the Federal Reserve Act itself. The U.S. had formally gone off the gold standard that year. Where before any member of the public could have walked into any Treasury office or Federal Reserve district bank and asked to have their banknotes redeemed with gold, neither institution was obligated to uphold this promise anymore. As of 1934, the Treasury would only buy gold from miners and other central banks at a rate of $35 per ounce—but this obligation didn't show up on the face of a Federal Reserve note, nor in the Federal Reserve Act. The promise to purchase gold at $35 was an informal one, with President Roosevelt remarking that the price "may be changed by the Secretary of the Treasury at any time without notice."

In addition to removing the gold redemption promise from the face of a Federal Reserve note, the 1934 series included a new feature: Federal Reserve notes were now legal tender for all debts public and private. It may seem strange to us now, but for the first twenty years of the Fed's existence, Federal Reserve notes could not legally discharge a debt. If James owed Judith $20, Judith could refuse to accept Federal Reserve notes from James, asking for something else instead, say United States notes which were legal tender. Thanks to the 1933 Thomas Amendment (pdf), Judith was now obligated to accept James's Fed notes as payment for the debt. I can only speculate on why Federal Reserve notes weren't originally made legal tender, but one reason is probably due to the memories of the inflation in the 1860s caused by legal tender greenbacks. If something isn't granted legal tender status, it can't do as much damage to the price level.

For almost thirty years nothing changed on the face of Federal Reserve notes until 1963 when the redeemable in lawful money promise was dropped, leaving only the stipulation that a note was legal tender for all debts, public and private.

By 1963, America's mongrel currency was pretty much a thing of the past. Ever since 1934 it had been illegal for gold certificates to circulate publicly. As for National bank notes, they had begun to be retired in 1935. The effort to remove silver certificates in denominations of $5 and above was initiated by John F. Kennedy in 1961. Getting rid of the $1 silver certificate was a bit more tricky. Believe it or not, but at the time there was no such thing as a $1 Federal Reserve note. For decades the nation's entire demand for $1 notes had been met solely by the Treasury's silver certificates. However, in 1963 the Fed finally debuted its first $1 note, upon which the Treasury began to cancel $1 silver certificates.

With most of the government's parallel currencies retired, the promise to redeem Federal Reserve notes in lawful money probably seemed pointless, if not confusing. Thus it no longer shows up on bills, despite being still encoded in section 16 of the Federal Reserve Act. In 2017, you can bring your note to a Federal Reserve for redemption in lawful money, but they will only give you another Fed note in return. The category of lawful money is meaningless.


Most members of the public don't know how the underlying monetary systems work, but they do see what is printed on its banknotes. To the public, the morphing set of promises on the face of a Federal Reserve note would have been one of the more visible manifestations of a shift from a mish mash of paper currencies issued by two different institutions and pegged to gold... to one single legal tender currency issued by the U.S.'s now-dominant monetary institution, the Fed—the Treasury receding into the background.

It's been over fifty years since the promise on Federal Reserve notes was last changed. What promises will be printed on U.S. money in the future? That's hard to say since we don't even know if paper money will be a part of the future. If the Fed were ever to update its currency by introducing a digital version to circulate along with its paper issue, those in charge of its design would have to think hard about the sorts of promises that will be granted to the bearer of those tokens. Would they be lawful money, tax receivable, and/or legal tender? These are questions that U.S. monetary authorities haven't had to ask themselves in a long time, not since the mongrel currency era when new money was introduced every decade or two.

Wednesday, October 18, 2017

An example of tax-driven money during the greenback era

Cartoon from 1864 poking fun at politicians and greenbacks (source and explanation)

During the greenback era, the Union government issued irredeemable paper money to help pay for its war against the Confederates. What many people don't realize is that there were actually two different strains of greenbacks—those printed before March 1862 and those printed after. Although these two strains had only slightly different properties, they were not fungible with each other and would go on to have drastically different values in the marketplace. Looking at the respective properties of each type gives some insights a thorny problem: why do colored bits of paper money have value?

One classic explanation for the value of fiat money is so-called 'tax-backing.' If the government stipulates that taxes must be paid using government-issued chits of paper, then that will be sufficient to give those chits a positive value. Back in 1910 economist Philip Wicksteed was one of the first economists to champion this explanation:
The Government, then, levying taxes upon the community, may say: "I shall take from you, in proportion to your resources, as a tribute to public expenses, the value of so much gold. You may pay it to me in actual metallic gold or you may pay it to me in anything which I choose to accept in lieu of the gold. If you do not give it me I shall take it from you, in gold or any other such articles as I can find, and which would serve my purpose, to the value of the gold. But if you can give me a piece of paper, of my own issue, to the face value of the gold that I am entitled to claim of you, I will accept that in payment." Now, as these demands of the Government are recurrent, there will always be a set of persons to whom the Government paper stamped with a unit weight of gold is actually equivalent to that weight of gold itself, because it will secure immunity from requisitions to the exact extent to which the gold would secure it. - The Common Sense of Political Economy, Book II Ch 7
The idea that a tax obligation can be the basis of money, or what Randall Wray has termed twintopt, is at the core of modern monetary theory, or MMT.

Let's see how the tax backing theory holds up during the greenback period. To set the context, South Carolina had seceded from the Union in December 1860, soon followed by ten other states. Hostilities between the Union and newly-formed Confederate states began in April 1861. To help fund the Union side of the war effort, an initial $50 million issue of greenbacks was authorized in July. This was to be the first government paper money emitted since the Second Bank of the United States had been wound up.

The act ruled that these notes were to be redeemable on demand in gold, a promise that was also inscribed on the face of each note (see below). They thus earned the nickname demand notes. This promise meant that, at the outset at least, their price could not deviate from par since any movement above or below their gold value would be arbitraged away. When redemption was rescinded a few months later on December 30, 1861, demand notes began to trade at a 1-2% discount to their face value in gold.

The second vintage of greenbacks, known as legal tender notes, was authorized two months later by the Legal Tender Act of February 25, 1862. This act provided for an issue of $150 million, much larger than the first vintage. One novelty is that the second batch of notes was declared legal tender, which meant that a creditor could not refuse to receive them at par in discharge of a debt. The legal tender property was extended to demand notes in March 1862. The second vintage of notes was also irredeemable, putting them on the same basis as the demand notes, which became irredeemable at the end of 1861.

What distinguished legal tender notes from demand notes? As the tweet above shows, the two vintages had visible differences—unlike demand notes, legal tender note did not have the promise to redeem on demand printed on their face. Demand notes also had an extra promise inscribed on them: "receivable in payment of all public dues". What this meant in practice is that the government accepted demand notes in payment for all taxes, including customs dues, whereas legal tender notes were only receivable in a narrow range of internal revenue taxes—and not for customs dues—which at the time made up the majority of Union tax revenues.

Receivability for customs dues was an important point of departure between demand notes and legal tender notes. Duties were priced in gold and could also be discharged with gold coins. So if an importer was on the hook for $x in custom duties, they could certainly scrounge up $x in gold coin to get rid of the obligation, but $x in demand notes would be sufficient to "secure immunity" from the tax. Not so with legal tender notes.

A given importer might only need a small portion of the demand notes in his possession to pay customs duties. Anything above that amount would be worth less than their face value to him, since gold—not demand notes—was necessary to buy goods internationally. However, if that importer could find other importers who were themselves under obligation to pay customs duties, and who would therefore value his remaining stash of demand notes for their tax receivability, then he might sell them his remaining demand notes at a price quite close to the value of gold coins.

So all that was needed to have irredeemable demand notes trade near the value of gold was a permanent market of tax payers who demanded those notes, and a flow of new notes that did not exceed the rate of drainage provided by the tax outlet. After all, if the supply of notes overwhelmed the amount of tax that needed to be paid, then notes would accumulate in importers pockets with no one willing to bid for them. Once everyone's taxes had all been paid up, demand notes would trade at a discount to gold coins. 

Tax receivability was successful in keeping the value of demand notes close to their gold value. The chart below shows the price of both demand notes and legal tender notes relative to gold through 1862-63.  Demand notes never fell to more than a 10¢ discount relative to gold coin. Calomiris blames this discount on the risk that receivability for customs duties would be revoked by the government before notes had been paid in.

Legal tender notes, however, fell to an ever larger discount relative to both gold coin and demand notes, reaching a 40¢ discount by March 1863. While legal tender notes were receivable for domestic taxes, these taxes did not account for a very large share of government revenues. Nor were these taxes priced in gold. Which meant that, unlike demand notes, legal tender notes were not benchmarked to some real good or price index.

So what, if anything, determined the price of legal tender notes? Here I'll introduce another theory for the value of money; the metallist viewpoint. Rather than tax receivability driving a currency's value, a metallist looks to the currency's intrinsic value. When banknotes are fully redeemable, their intrinsic value is determined by the underlying gold on which the note is a claim, the value of which is set in the market for precious metals in technology and the arts.

In the case of legal tender notes, which were no longer redeemable, the realization of intrinsic value had only been delayed to some future point in time when gold convertibility would be re-adopted. This eventual re-mooring date was in turn a function of the Union government's ability to win the war, among other factors. According to this theory, the steady decline in the gold value of legal tender notes in the chart above can be blamed on the realization by the public that the war would last much longer than most originally thought, pushing the re-mooring date ever further into the future.

While the 1861 issue of demand notes had all been bought back and cancelled by the government by mid-1863, greenbacks remained outstanding even after the war had been won. Their discount to gold continued to widen till July 1864 at which point their price steadily rose. See the chart below. The steady return to par probably can't be explained by the tax-backing theory. Improved odds that the government's fiscal situation would allow it to resume gold convertibility—an event that finally occurred in 1879—is a better explanation. There have been a few interesting accounts written about the value of greenbacks over the full 1862-1879 period, including Wesley Clair Mitchell's A History of the Greenbacks in 1903, but also more recent contributions here (Calomiris), here (Smith & Smith) and here (Willard et al).

In sum, the U.S. government was issuing three non-fungible currencies by 1862. Coins and legal tender notes operated under the principles of a metallic money whereas the third, demand notes, seems to have been a purely tax-driven money as described by Wicksteed. So what about a modern dollar note? Is a Federal Reserve note like an 1861 demand note and mostly tax-driven, or like legal-tender notes and operating on a metallic basis?

I'm not entirely sure, but my guess is that it is a messy combination of both. When it comes to money, I'm not a believer in any one theory. Although the odds of a future return to the old 1972 gold redemption rule of 0.024 ounces per dollar is non-existent, the metallic explanation for the dollar's value continues to be relevant. Instead of redeeming currency with fixed amounts of metal as in days of yore, modern central banks repurchase notes with financial assets held in their vault in a manner that is consistent with hitting an inflation target. This is very much like 1800s-style metallism, except with an ever-shrinking CPI basket in the place of a fixed amount of gold.   


Wesley Clair Mitchell, A History of the Greenbacks

PS: I recently started a discussion board here. Feel free to bring up topics not covered on my more recent blog posts, suggest posts, or discuss ideas that appear on my Twitter feed. I don't like Twitter for long-form discussion; I'd much rather divert them to the board.

PPS: Nathan Tankus responds here.

Thursday, October 12, 2017

The ubiquitous Spanish dollar—a photo essay

"The head of a fool on the neck of an ass."

That's how Londoners described the strange silver coin pictured above, which first appeared in Britain in 1797. Due to worries that Napoleon was about to invade the British Isles, a run had developed on the Bank of England. In response, Parliament allowed the Bank to refuse to redeem its notes with gold coins, but this had only resulted in an inconvenient shortage of coins.

To remedy the shortage, the Bank of England decided to open its vault and put its hoard of silver coins into circulation. Complicating matters was the fact that these coins were not native shillings or pennies, but Spanish dollars, otherwise known as eight real pieces. As such, they had to be re-purposed into local currency. The Bank of England accomplished this task by stamping the head of King George III—the fool—on the neck of Charles IV of Spain—the ass—who occupied the obverse side of that era's version of the Spanish dollar. The Bank then declared that all stamped Spanish dollars were to be worth four shillings and nine pence. People flocked to the Bank of England's window to get their hands on the new coins.  

That the Spanish dollar temporarily became part of England's circulating media of exchange was due to its ubiquity—in its heyday the Spanish dollar, like today's U.S. dollar, was everywhere. After discovering huge amounts of silver in the new world, the Spaniards had set up a mint in Mexico City in 1536 to produce large quantities of silver coins. Mints in Potosi (Bolivia), Lima (Peru), and elsewhere soon followed. It is estimated that some four-fifths of the world's silver produced between 1493 and 1850 came from Spanish America, almost all of it in the form of Spanish dollars! 

Although their design changed over the centuries, by the 1700s the Spanish dollar—easily identified by the two "pillars of Hercules" on the reverse side of the coin—had become known and accepted all over the world. As the map above shows, the two pillars refer to the Rock of Gibraltar in Spain and its counterpart on the Moroccan side of the narrow strait separating Africa from Europe.

In addition to England, the Spanish dollar shows up in the Caribbean islands where it made up a major part of the local coinage. To provide small change, locals cut dollars up into pieces. Below, an 1806 dollar used in Guadeloupe has had its centre carved out of it and stamped with the letter "G". Earmarking of coins was done to prevent exportation and restrict usage within borders. In this way, local currencies were literally piggybacked into existence off of the ubiquitous Spanish dollar.

Source: The Spanish Dollar as Adapted for Currency in Our West Indian Colonies

In the next image, a Trinidadian version of the dollar has been carved up like a pizza. The half piece has been countermarked with a "6", the third with a "4", and the sixth with a "2". These numbers refer to the number of bits or bitts the piece was worth, the bit being a local accounting unit. By marking the letters "TR" on each dollar to indicate Trinidad, the Spanish dollar has been forked to create a second currency.

Spanish dollar marked with TR for Trinidad

All possible methods of cutting up Spanish dollars were used. Here's a 1797 dollar from Saint Lucia that has been cut up lengthwise, like a loaf of bread, with "SLucie" being countermarked on each section.

Saint Lucia Spanish dollar

The Spanish dollar also made a notable appearance on the opposite side of the globe. In 1813, the colony of Australia minted its first currency, the so-called holey dollar. To deal with a shortage of coins, the governor of New South Wales imported thousands of Spanish dollars. A convicted forger was contracted to cut the centre from each one, countermarking each of the resulting pieces with the name of the colony and its nominal value in shillings and pence. As in the case of the Caribbean Islands, this mutilation of the dollar was an attempt to render them useless outside of the colony. The outer ring was rated at five shillings and the centre—the dump—at fifteen pence, or one shilling and three pence. 

The Spanish dollar, which would become the Mexican dollar after Mexico won its independence from Spain, was also popular in China. Merchants and professional money exchangers, or shroffs, would test a coin to verify its silver content. According to James Gullberg, they went about this by balancing the coin on their finger and tapping it. If it rang, it was legitimate. The theory goes that once the coin had passed their purity test, a shroff would stamp that coin with his own peculiar chopmark—a Chinese character, an emblem, symbol, or a pseudo character. (I discussed this practice in more depth here). The more chops a Spanish dollar had, the better its quality—even after it had been chopped beyond recognition.

1807 Spanish dollar with chopmarks

Pictured above is a chopped Spanish dollar from 1807. According to Gullberg, Chinese merchants preferred the Carolus IIII dollar, which was issued between 1772-1810. They referred to it as four work, the IIII resembling the Chinese character for going to work. The influence of the Spanish/Mexican dollar continued even into the 19th century: the banknote below, issued in 1912 by the Sino-Belgian Bank, is redeemable in Mexican dollars.

Moving back west, the Spanish/Mexican dollar had a key role to play in the North American economy up to the mid-1800s. In both Canada and the U.S., the unit of account function of money was separated from the medium of exchange function. Early colonists kept prices in pounds, shillings, and pence unit of account, but Spanish dollars and their subdivisions were used to transact business. Because of the long distance from England, there simply weren't enough shillings and pennies to make do.

Even though a local U.S. dollar coin—which was modeled off of the Spanish dollar—had been minted as early as 1794, the Spanish version remained by far the most popular form of coinage in the U.S. The presence of the Spanish dollar was even enshrined in American law, the coins having been declared legal tender in 1793. Many people have even speculated that the famous dollar sign, $, has been bastardized from the symbol for peso.

Source: Evolution of the Dollar Mark (1912)

To illustrate the ubiquity of the Spanish dollar in the U.S., below I've included a 6¼ cent note issued in 1816 by Easton and Wilkesbarre Turnpike Company, which built and operated private toll roads. This odd denomination only makes sense if you consider it in the same context as the Spanish dollar. Unlike U.S. dollars, which were divisible by 100 into cents, the Spanish dollar was divisible into 8 reals. A one-real coin, a fairly common unit around that time, would have been worth 12½ cents, and a ½ real coin worth 6¼ cents. So while the turnpike company's decision to issue a 6¼ cent note might seem odd to us today, it was probably meant as a convenience to its users, who would have been more familiar with Spanish coins than any other type of coin.

As for Canada, below is an 1819 $5 bill from the Bank of Upper Canada. To help customers who couldn't read, the bank conveniently printed an image of five Spanish dollars at the bottom of the note.

This 1819 banknote has five Spanish dollars illustrated along its bottom

For a number of reasons, the Mexican dollar was eventually killed off . The U.S. had originally been on a bimetallic standard, but in 1834 the authorities changed the silver-to-gold ratio to 16:1 from 15:1, in the process slightly overvaluing gold. The discover of the yellow metal in California only further pressured the price of gold down relative to silver, exacerbating the legally-imposed undervaluation of silver. As a result, silver coins began to disappear. After all, anyone who had a stash of silver coins would have found it more profitable to melt them down and export them overseas where they traded at their true economic value. The U.S. had effectively flipped onto a gold standard. To compound matters, Spanish dollars lost legal tender status in 1856, which would have further crimped their demand.

As for Asia, several competing coins were created, including the Japanese yen, a near-replica of the Mexican dollar. The U.S. began to produce special trade dollars that were to be used solely in Asian trade (I wrote about the trade dollar here). Further reducing the demand for Mexican silver was the decision by many nations to follow the major western nations and officially adopt gold standards, including Japan in 1897 and the Philippines in 1904.


Even though it was eventually eclipsed, there are still modern day echoes of the old Spanish dollar. The pejorative term two bit is still used in the U.S. as a synonym for cheap or insignificant, as in two-bit thief. One bit was ⅛ of a dollar, or 12.5¢, and two-bits 25¢, so a two-bit thief is a 25¢ thief. Amazingly, the practice of pricing in sixteenths continued on the New York Stock Exchange up till 2001. The Toronto Stock Exchange decimalized in 1996.

Up here in Canada, Quebecers use the word piasse for dollar. I had always assumed that they meant piece, but I recently discovered that the wording actually goes back centuries to piastre, the French word for the Spanish peso. Even more interesting is the vernacular usage of trente-sous (or 30 sous) to mean 25-cents, which I only understood after reading this from Frédéric Farid.
After the '87 market crash, a business man begs for 25¢ (Source)

The story goes like this. In the 18th and 19th Centuries, an English halfpenny was referred to in the colony of Lower Canada (i.e. Quebec) as a sou. The pound-shilling-pence unit of account (£/s/d) was still in use at the time but Spanish dollars were the most common coin in circulation in both Upper and Lower Canada. Dollars were officially rated by the British authorities at $4 for each £1. A quarter Spanish dollar (a two-real coin or 25 cent piece) would therefore be worth £0.0625.

Since an English pound contained 240 pence, that meant a quarter Spanish dollar was worth 15 English pennies, or 30 half-pennies. And given that Quebecers referred to half-pennies as sous, that gave rise to the ongoing practice of referring to 25 cents as 30 sous. When they use this term today, Quebecers are just referring to the archaic exchange rate between Spanish dollars and pence.

Source: La Maison nationale des Patriotes

This odd system of conversion really hits home when you work through the above scrip note emitted by Distillerie St-Denis, printed in 1837.  It is denominated in three different units. Trente-sous appears twice in French and once as XXX, while in the English paragraph near the bottom the £/s/d equivalent of one shilling and three pence, or fifteen English pennies, appears. Finally, a two-real coin, or quarter dollar—the pillars of Hercules side showing—has been emblazoned smack in the centre.

To end things off, you can also see a ghost of the old Spanish dollar in Venezuela's 12½ centimos coin. This coin is an eight of a bolivar—reckoning in terms of eights and sixteenths is a hallmark of the days of the old pillar coin.
Given Venezuela's ongoing hyperinflation, the likely destiny of the 12½ centimos coin is to be nothing more than another two-bit dead coin.

George Selgin, Good Money
A. Piatt Andrew, The End of the Mexican Dollar (link)
J.B. Caldecott. The Spanish Dollar as Adapted for Currency in Our West Indian Colonies (link)

Wednesday, September 27, 2017

The siren call of T+0, or real-time settlement

The NYSE's clearinghouse in 1898, six years after its founding

Traditional financial systems often get mocked for being slow. In North America, for instance, securities markets have recently switched from T+3 to T+2 settlement. Before, if you sold a stock the cash would only appear in your account three days after the trade—now settlement has been moved to a blazing fast two days. In an age where mail is transmitted in milliseconds, this delay seems terribly old fashioned. Or take automatic clearing house (ACH) payments in the U.S. Earlier this month the ability to make same-day ACH debit payments was rolled out, an improvement over the three or four days they used to take, but still no where near immediate.

The snail-like pace of securities and ACH settlement is often contrasted to real-time settlement, say like how payments using banknotes, coins, or bitcoins are finalized the moment the token leaves ones wallet and enters the destination wallet. Or take real-time gross settlement systems operated by central banks, over which a transfer of balances from one account to the other occurs instantaneously and is irrevocable. In the case of securities settlement, why only go from T+3 to T+2? Why not go straight to T+0?

Don't be beguiled by settlement speed. Slow isn't necessarily a bug—it's often a feature. Imagine the following scenario. You and your friends play poker every day at a cafe. To buy into the game, cash or bitcoins are required. And at the end of each game, cash or bitcoin is paid out to the winners. The problem with this system is that each day all players have to lug a transactions medium to the cafe and back from it—and this involves a sacrifice. Banknotes and coins take up lots of space and can be easily stolen. Like bitcoin, they don't yield interest—so a stream of interest income is being foregone to play poker. Once the game is done, the laborious process of counting out cash and banknotes occurs. In the case of bitcoin, the payouts are costly since each one involves incurring a fee to send bitcoins from one wallet to another.

Participants in financial markets have adopted a time-tested strategy to avoid much of the work involved in repetitive use of transactions media like cash: substitute them with IOUs that are only settled from time-to-time. Returning to the poker example, rather than stumping up cash each game players can buy-in using IOUs denominated in cash or bitcoin. These IOUs are recorded in a ledger. Rather than cashing out at the end of the game each player's balance is held over to the next day, only to be updated subject to that day's results. These ledger balances continue to be updated at the close of each game until after (say) the tenth game, everyone finally decides to settle their accounts, upon which all debtors, or losers, bring cash and/or bitcoin to the cafe to pay off winners, or creditors. The system has settled—not in real-time—but T+10.

The advantage of delayed settlement is that quid pro quo is achieved with one set of transactions conducted at the end of the 10-game cycle rather than a set of transaction for each game. No more tedious counting out coins each day or daily bitcoin fees. Because the obligation to carry around cash is kept to a minimum, interest income needn't be sacrificed by players. Nor are there any nuisances of storing cash.

While slowing down the system reduces the amount of work that must be done, it comes at the expense of flexibility and safety. One of the benefits of a cash payment is that transactions media are immediately available for use in subsequent transactions. If a player can only get cash out of the game after ten games, they will have to be sure that they don't need that cash for other payments in the interim. Slowing down the system also introduces credit risk into the system. Players may be unable or unwilling to honor their IOUs at the end of the cycle. Lengthening the cycle only increases the odds of settlement failure due to insolvency or bankruptcy.

If the benefits that your friends expect to harvest from delaying poker settlement—the reduction in work and fees—outweigh the aforementioned costs, then a T+10 system makes a lot of sense. Keep this in mind whenever you encounter a real-life financial transaction like an ACH payment that takes a long time to settle. The system's sluggishness may be designed that way because the conservation of work and transactions costs outweighs the inconveniences of not having immediate availability of transactions media and exposure to credit risk.

Even the bitcoin ecosystem has adopted various forms of delayed settlement. Thanks to high fees and long wait times, Bitcoin companies have been avoiding direct exchanges of bitcoins among each other in favor of netting out bitcoin-denominated IOUs, says Izabella Kaminska, only settling net amounts after some time has passed. And Bitcoin developers are working towards introducing something called the Lightning Network, which will allow users to make payments using fully-backed bitcoin depository receipts rather than having to settle trades directly on the blockchain using regular (and peskily-slow) bitcoins. 

Let's finish off by revisiting the recent shift from T+3 to T+2 securities settlement in the US and Canada. Interestingly, if you zoom out you'll see that over time the New York Stock Exchange shows a predominant tendency to lengthen the settlement cycle, not shorten it. The recent move to T+2 only brings the exchange back to the same settlement speed at which it operated at from 1933 to 1952. Prior to 1933, next-day settlement, or T+1, had been standard.

The lengthening of the cycle to T+2 in 1933, which corresponded with an increase in stock trading volumes, was implemented to "ease the work" of brokerage clerks. Back then all securities were recorded in physical form, so settlement required the transportation by hand of certificates from one office to another by an army of runners. In the face of growing trade volumes, the only way to maintain T+1 settlement would have been to do much more work, which meant hiring more clerks and runners—costs that would be offloaded onto clients. Slowing down the system to T+2 from T+1 presumably would have kept things cheap.

The 1968 switch to T+5 settlement was an effort to cope with the famous "back office crisis." Investors who were too young to remember the Great Depression had begun to arrive in droves to equity markets in the early-60s. Back office clerks could not keep up with amount of work required to settle trades. At one point the NYSE even closed on Wednesdays to help deal with the backlog.

The recent move back to T+2 means that cash will appear in investors' accounts 24 hours earlier, making it easier for them to meet subsequent payments deadlines. Credit risk is reduced too. Brokers conduct trades with other brokers on behalf of their clients, building up credit and debits over the settlement cycle. The shorter the cycle, the quicker these debts will be unwound by transfers of stock and cash, the resulting savings hopefully flowing through to customers.

Why not go straight to real-time, or T+0? The move from T+3 to T+2 means one less day over which brokers can 'net out' their respective debits and credits so as to conserve on transactions costs. T+0 means no netting-out window whatsoever—and that would impose a terrific amount of work on the system. Like I say, it's a trade-off. Real-time settlement is no panacea.

P.S. Here's a great article on the history of securities clearing and settlement: Was Trade Settlement Always on T+3? A History of Clearing and Settlement Changes, by Kenneth Levine (1996)

Wednesday, September 20, 2017

Post-demonetization, the role of cash has been diminished

Narendra Modi's demonetization scheme, which involved the sudden cancellation of all ₹1000 and ₹500 rupee notes, was designed to hurt those in the underground economy who had stashed away large amounts of cash. But many smaller cash users got caught up in the blast radius. Has Indian behaviour surrounding cash changed? More specifically, what long-term impact has the painful demonetization process had, if any, on the population's preferences for holding cash?

Luckily for us, we have a great data set for evaluating this question: the supply of currency in circulation. If the propensity of Indians to hoard notes has changed, we'd expect to see a long-term reduction in currency outstanding relative to trend.

Before we take a look at the data, I want to make one point on the economics of paper money. The old phrase "cranking up the printing presses" is a bad one because it implies that central bankers play an active role in deciding the quantity of paper money in the economy—they don't, their role is purely passive. At any point in time the stock of paper money in circulation has been determined by the decisions of the public. If people feel they need more notes, they'll go to their bank and withdraw them, the bank in turn submitting a request to the central bank to provide more. The job of a central bank is a simple one: react to the public's desire by printing new notes, or—in the case of a decline in demand for paper currency—withdrawing them.

While the stock of notes is always a reflection of people's preferences, there are a few exceptions. If there is a sudden spike in the demand for banknotes, and the central bank's printers can't keep pace, the public's desire for more paper will be temporarily frustrated.

This frustration is exactly what occurred in India in the months after demonetization. A whopping 85% of the nation's notes were suddenly cancelled. Indians needed to get their hands on replacement notes in order to move back to their preferred holdings of cash, but the RBI didn't have the printing capacity to meet this demand. Lineups at banks and ATMs grew, and withdrawal limits were imposed, a sure indication that the quantity of paper rupees in circulation was not consistent with Indian's preferences.

With 315 days having passed since demonetization was announced, has the RBI finally caught up to the public's desired stock of cash? I'd argue that it has. If we go back to this post by James Wilson, he calculates that—given various assumptions about the number of presses brought on line and shifts per day—all demonetized ₹500 would be replaced by as early as mid-June and late as mid-September 2017. We have now passed the later of the two dates.

The second reason that I think the RBI has caught up to Indians' demand for cash is the return of the same sawtooth pattern in the weekly currency-in-circulation data that characterized the pre-demonetization period. See below:

While currency in circulation consistently grew from mid-December 2016 onward, consistent with a dearth of paper money, it finally had its first week-to-week drawdown in July 2017, followed by several since then. The Indian public would only have returned currency to their banks if their collective demand for currency was finally satiated. If there was still a structural deficiency in the supply of notes, they would have kept these notes in their wallets rather than bringing them in.

The point of all this is to show that current data on the stock of rupees represents the true preferences of Indians, not RBI printing constraints. So let's take a closer look at the data itself. Prior to demonetization, India's currency supply had been consistently growing at around 15% a year, as represented by the black trendline below. In September 2016, just two months before demonetization, Indians had collectively chosen to hold ₹17.5 trillion in cash (around US$272 billion).

If there has been no change in cash-holding behaviour then the pre-demonetization growth path of 15% should still be in effect. If so, from a post-demonetization low Indians would have rebuilt their cash holdings to a trend amount of ₹20 trillion today. Instead, currency in circulation clocks in at just ₹15.8 trillion, around ₹4.2 trillion (or US$65 billion) less then the counterfactual.

So demonetization has certainly modified the cash-using behaviour of Indians. Their fingers burned, they have decided to hold just 75% of what they would otherwise have in their wallets had demonetization never occurred. It remains to be seen if this is a one-time change in the level of currency in circulation or whether the 15% growth rate itself will be permanently modified. In other words, from this point on growth in cash can either continue to grow 15% from a lower base of ₹15.8 trillion, or a new rate of growth might emerge, say 5-10%. Demonetization will have been a more effective cudgel if it succeeded in changing the actual rate of growth, not just the level.

One wonders what filled the ₹4.2 trillion void. Has the cash-using public decided to hold a larger stock of state-sanctioned digital money (i.e. mobile money balances and bank accounts) or have they simply chosen to swap one form of anonymous "black money" for another (i.e. cash for gold, diamonds, real estate)? I've gathered some data showing that at least some of the gap is being filled by the former.

The above chart shows that the number of point-of-sale (POS) terminals installed has experienced a one-time shot to the arm after last fall's demonetization announcement.

The other two charts show that the value of POS transactions using debit and credit cards moved to a new and higher plateau, as did mobile wallet payments (like Paytm). It is notable that transactions using these methods are not returning to their previous levels (although this is from a relatively small base) indicating that these payments options weren't mere fill-ins during the cash shortage period, but have enjoyed a permanent increase in adoption. Given that these new forms of digital payment are doing much more "work" as media of exchange, it probably makes sense to assume that they have captured at least some of the substitution out of cash.

So demonetization has certainly changed Indians' attitudes to cash. They like the stuff a bit less than before and are holding fewer bills in their wallets. This change might seem like an inevitable one, but at the time demonetization was announced it was by no means certain that it would have any effect whatsoever on tastes for paper money. Whether these effects justify the whole affair is a much more complex question.

P.S. You can also tell a story in which demonetization caused a depression in India's informal economy, and since economic activity in this sector plunged people had less demand for notes. So contrary to everything I wrote above, a one time change in preferences for alternative forms of money over cash didn't occur--people simply need less of the stuff because there are fewer transactions to be made. I don't doubt that demonetization hurt the informal economy, but I am skeptical that it would have led to a 25% reduction in cash usage. I think the changing-of-tastes story explains the data better.

Wednesday, September 13, 2017

When a rising stock market is a bad thing

If the world had a single cauldron for mixing various monetary phenomena, it would be Zimbabwe. Over the last two decades, it has experienced pretty much everything that can happen to money, from hyperinflation to deflation, demonetization to remonetization, dollarization and de-dollarization, bank runs, bank walks, and more.

Adding to this mix, the Zimbabwe Industrial Index—an indicator of local stock prices—has recently gone parabolic, having more than tripled over the last twelve months. That's a good sign, right? Beware, these gains aren't real. As is often the case in Zimbabwe, the rise in stock prices is a purely monetary phenomenon.

Ever since the great Zimbabwean hyperinflation led to the domestic currency becoming worthless in 2008, U.S. dollars have served as the nation's currency and unit of account. However, Zimbabwe's central bank, the Reserve Bank of Zimbabwe (RBZ), has spent much of the last two or so years surreptitiously bringing a new parallel monetary unit into circulation. These new units, informally referred to as RTGS dollars, are a digital form of money, specifically a deposit held at the central bank. (I described them here).

At the outset, RTGS dollars were denominated in U.S. dollars and supposedly convertible into genuine U.S. banknotes. We now know these units were only masquerading as U.S. dollars. By early 2016, huge lineups began to appear outside banks as Zimbabweans unsuccessfully tried to convert their deposits into real U.S. cash. When conversions finally became possible it was only because the RBZ had introduced its own paper currency, a 'bond note', in late 2016. These bond notes were themselves supposed to be fully fungible with U.S. dollars thanks to a promise of 1:1 convertibility, at least if you believed the nation's central banker, but this has never been the case.   

Over the last year a great redenomination has been occurring as all Zimbabwean prices—including that of ZSE-listed stocks—are shifted over from a genuine U.S. dollar standard to an RTGS dollars/bond note standard. Prior to 2016, if you sold a stock or received a dividend, you'd get U.S. dollars, or at least a pretty decent claim on the real thing. Now, you get RTGS dollars—which can only be cashed out into an equally dodgy parallel currency, bond notes.

The incredible 300% rise in the Zimbabwe Industrial Index is a reflection of the redenomination of stocks into an inferior monetary unit and that unit's continued deterioration. For instance, if you were willing to sell your shares of Delta Brewing for $10 prior to the redenomination, you'd only be willing to sell them at a much higher level post-redenomination, say $15, in order to adjust for the diminished purchasing power of the money you'll receive upon sale. RTGS dollars are not U.S. dollars. After adjusting the Zimbabwe Industrial Index for the decline in the value of the money in which stocks are denominated, things certainly wouldn't look as bullish as the chart above indicates.

What is the actual exchange rate between RTGS dollars/bond notes? We can get a pretty good idea by looking at the prices of dual-listed shares. The shares of Old Mutual, a global financial company, trade in both Harare and on the London Stock Exchange. See chart below, courtesy of Gareth on Twitter.

Old Mutual investors have the ability to deregister their shares from one exchange and transfer them for re-registration on the other. Arbitrage should keep the prices of each listing in line. After all, if the price in London is too high, investors need only buy the shares in Zimbabwe, transfer them to London, sell, and repurchase in Zimbabwe, earning risk-free profits. If the price in Zimbabwe is too high, just do the reverse.

The ratio of the two Old Mutual listings (the green line above) provides us with the implicit exchange rate between genuine U.S. dollars and dollars held in Zimbabwe. In 2009 the two listings traded close to parity (i.e. around $2 each), which makes sense because Zimbabwe had dollarized by then, and dollars-in-Zimbabwe were fungible with regular dollars. From 2010 to 2016 the dollar-denominated price in London was above the price in Zimbabwe. This discrepancy may be due in part to the fact that Harare-listed Old Mutual shares aren't very liquid, so they suffer a liquidity discount. Another reason is that the authorities place a ceiling (i.e. fungibility limits) on the number of Old Mutual shares that can be deregistered from the Harare market and dropshipped into London. With all of the space under the ceiling having presumably been used up, it would have been impossible to arbitrage the difference between the two prices, the Harare counter falling to a permanent discount.

So the Old Mutual ratio was probably not a good indicator of the implicit exchange rate between 2010 and 2016. However, in June 2016 this ceiling was raised, at which point arbitrage would have once again been possible. As such, the ratio would have probably returned to providing a decent indicator of the exchange rate between a dollar-in-Zimbabwe and a genuine dollar. 

You can see that Old Mutual is currently valued at $5.80 in Zimbabwe whereas it only trades in London for around $2.66 per share (after converting from pounds into US dollars). This means that Zimbabweans are willing to put $5.80 in one end of the sausage maker in order to get $2.66 out from the other. So a dollar-in-Zimbabwe, which was trading at par to U.S. dollars just two years ago, is now worth just 46 cents. That's quite the inflation rate. I don't see things slowing down, either.

P.S.: Some investors will no doubt want to say the same thing is going on with the US and Europe with loose monetary policy creating so-called asset price inflation. I disagree.

P.P.S: Gareth has provided another chart. It shows the implied exchange rate between dollars-in-Zimbabwe and genuine U.S. dollars using the only other fully transferable dual-listed counter, PPC, a cement company that is also listed in Johannesburg. Note how close the PPC rate follows the Old Mutual rate.