Friday, May 27, 2016

From ancient electrum to modern currency baskets (with a quick detour through symmetallism)

Electrum coins [source]

First proposed by economist Alfred Marshall in the late 19th century as an alternative metallic standard to the gold, silver and bimetallic standards, symmetallism was widely debated at the time but never adopted. Marshall's idea amounted to fusing together fixed quantities of silver and gold in the same coin rather than striking separate gold and/or silver coins. Symmetallism is actually one of the world's oldest monetary standards. In the seventh century B.C., the kingdom of Lydia struck the first coins out of electrum, a naturally occurring mix of gold and silver. Electrum coins are captured in the above photo.

While symmetallism is an archaic concept, it has at least some relevance to today's world. Modern currencies that are pegged to the dollar (like the Hong Kong dollar) act very much like currencies on a gold standard, the dollar filling in for the role of gold. A shift from a dollar peg to one involving a basket of other currencies amounts to the adoption of a modern version of Marshall's symmetallic standard, the euro/yen/etc playing the role of electrum.

The most recent of these shifts has occurred with China, which late last year said it would be measuring the renminbi against a trade-weighted basket of 13 currencies rather than just the U.S. dollar. Thus many of the same issues that were at stake back at the turn of the 19th century when Marshall dreamt up the idea of symmetallism are relevant today.

So what exactly is symmetallism? In the late 1800s, the dominant monetary debate concerned the relative merits of the gold standard and its alternatives, the best known of which was a bimetallic standard. The western world, which was mostly on a gold standard back then, had experienced a steady deflation in prices since 1875. This "cross of gold" was damaging to debtors; they had to settle with a higher real quantity of currency. The reintroduction of silver as legal tender would mean that debts could be paid off with a lower real amount of resources. No wonder the debtor class was a strong proponent of bimetallism.

There was more to the debate than mere class interests. As long as prices and wages were rigid, insufficient supplies of gold in the face of strong gold demand might aggravate business cycle downturns. For this reason, leading economists of the day like Alfred Marshall, Leon Walras, and Irving Fisher mostly agreed that a bimetallic standard was superior to either a silver standard or a gold standard. (And a hundred or so years later, Milton Friedman would come to the same conclusion.)

The advantage of a bimetallic standard is that the price level is held hostage to not just one precious metal but two; silver and gold. This means that bimetallism is likely to be less fickle than a monometallic standard. As Irving Fisher said: "Bimetallism spreads the effect of any single fluctuation over the combined gold and silver markets."  Thus if the late 1800s standard had been moved from a gold basis to a bimetallic one, the stock of monetary material would have grown to include silver, thus 'venting' deflationary pressures.

Despite these benefits, everyone admitted that classical bimetallism had a major weakness; eventually it ran into Gresham's law. Under bimetallism, the mint advertised how many coins that it would fabricate out of pound of silver or gold, in effect setting a rate between the two metals. If the mint's rate differed too much from the market rate, no one would bring the undervalued metal (say silver) to the mint, preferring to hoard it or export it overseas where it was properly valued. The result would be small denomination silver coin shortages, which complicated trade. What had started out as a bimetallic standard thus degenerated into an unofficial gold standard (or a silver one) so that once again the nation's price level was held hostage to just one metal.

The genius of Alfred Marshall's symmetallic standard was that it salvaged the benefits of a bimetallic standard from Gresham's law. Instead of defining the pound as either a fixed quantity of gold or silver, the pound was to be defined as a fixed quantity of gold twinned with a fixed quantity of silver, or as electrum. Thus a £1 note or token coin would be exchangeable at the Bank of England not for, say, 113 grains of gold, but for 56 grains of gold together with twenty or so times as many grains of silver. The number of silver and gold grains in each pound would be fixed indefinitely when the standard was introduced.

Because symmetallism fuses gold and silver into super-commodity, the monetary authority no longer sets the price ratio between the two metals. Gresham's law, which afflicts any bimetallic system when one of the two metals is artificially undervalued, was no longer free to operate. At the same time, the quantity of metal recruited into monetary purposes was much larger and more diverse than under a monometallic standard, thus reducing the effect of fluctuations in the precious metals market on aggregate demand.

While symmetallism was an elegant solution, Alfred Marshall was lukewarm to his own idea, noting that "it is with great diffidence that I suggest an alternative bimetallic scheme." To achieve a stable price level, Marshall preferred a complete separation of the unit of account, the pound, from the media of exchange, notes and coins. This was called a tabular standard, a system earlier proposed by William Stanley Jevons. The idea went nowhere, however; the only nation I know that has implemented such a standard is Chile. As for Fisher, he proposed his own compensated dollar standard plan, which I described here.

The urgency to adopt a new standard diminished as gold discoveries in South Africa and the Yukon spurred production higher, thus reducing deflationary pressures. None of these exotic plans—Marshall's symmetallism, Jevons tabular standard, or Fisher's compensated dollar—would ever be adopted. Rather, the world kept on limping forward under various forms of the gold standard. This standard would be progressively modified through the years in order to conserve on the necessity for gold, first by removing gold coin from circulation and substituting convertibility into gold bars (a gold bullion standard) and then having one (or two) nations take on the task of maintaining gold convertibility while the remaining nations pegged to that nation's currency (a gold exchange standard).


Let's bring this back to the present. In the same way that conditions in the gold market caused deflation among gold standard countries in the late 1800s, the huge rise in the U.S. dollar over the last few years has tightened monetary conditions in all those nations that peg their currency to the dollar. To cope, many of these countries have devalued their currencies, a development that Lars Christensen has called an 'unraveling of the dollar bloc.'

A more lasting alternative to re-rating a U.S. dollar peg might be to create a fiat version of electrum; mix the U.S. dollar with other currencies like the euro and yen to create a currency basket and peg to this basket. China, which has been the most important member of the dollar bloc, has turned to the modern version of symmetallism by placing less emphasis on pegging to the U.S. dollar and more emphasis on measuring the yuan against a trade-weighted basket of currencies. This means that where before China had a strictly made-in-the U.S. monetary policy, its price level is now determined by more diverse forces. Better to put your eggs in two or three baskets than just one.

Bahrain, Oman, Qatar, Saudi Arabia and United Arab Emirates are also members of the dollar bloc. Kuwait, however, links its dinar to a basket of currencies, a policy it adopted in 2007 to cope with the inflationary fallout from the weakening U.S. dollar. In an FT article from April entitled Kuwaiti currency basket yield benefits, the point is made that Kuwait has enjoyed a more flexible monetary policy than its neighbours over the recent period of U.S. dollar strength. Look for the other GCC countries to mull over Kuwaiti-style electrum if the U.S. dollar, currently in holding pattern, starts to rise again.

Modern day electrum can get downright exotic. Jeffrey Frankel, for instance, has suggested including commodities among the basket of fiat currencies, specifically oil in the case of the GCC nations. Such a basket would allow oil producing countries to better weather commodity shocks than if they remained on their dollar pegs. If you want to pursue these ideas further, wander over to Lars Christensen's blog where Frankel's peg the export price plan is a regular subject of conversation.

Friday, May 20, 2016

Those new Japanese safety deposit boxes must all be empty

Remember all the hoopla about Japanese buying safety deposit boxes to hold cash in response to the Bank of Japan's decision to set negative rates? Here is the Wall Street Journal:
Look no further than Japan’s hardware stores for a worrying new sign that consumers are hoarding cash--the opposite of what the Bank of Japan had hoped when it recently introduced negative interest rates. Signs are emerging of higher demand for safes—a place where the interest rate on cash is always zero, no matter what the central bank does.
Well, three month's worth of data shows no evidence of unusual cash demand. As the chart below illustrates, the rate at which the Bank of Japan is printing the ¥10,000 note shows no discontinuity from its pre-negative rate rise. In fact, demand for the ¥10,000 is far below what it was in the 1990s, when interest rates were positive.

I should remind readers that the Bank of Japan, like any central bank, doesn't determine the quantity of banknotes in circulation; rather, the public draws those notes into circulation by converting deposits into notes. So this data is a pure indicator of Japanese cash demand.

The lack of interest in switching into yen notes should come as no surprise given the experience of other nations that have set negative interest rates. Data from Sweden, Denmark, and Switzerland has consistently shown that it takes more than just a slight dip into negative territory before the dreaded "lower bound," the point at which the public converts all their deposits into banknotes, is encountered.

For instance, despite the setting of a -0.5% repo rate by Sweden's Riksbank, Swedish cash in circulation continues to decline. I won't bother to provide a chart, you can go see the data here.

As for the Danes, the growth rate in Danish cash demand continues to hover near its long term average of 3.7%/year. This despite the fact that the Danmarks Nationalbank, the nation's central bank, is setting a deposit rate of -0.65%. I've charted it out below:

Definitely no lower bound in Denmark, at least not yet.

Finally, we have Switzerland where the Swiss National Bank has maintained a -0.75% rate since December 2014. Back in February, the WSJ and Zero Hedge were making a fuss out of the sudden jump in the demand for the 1000 franc note, in effect blaming the build up on the lower bound. I wrote a rebuttal at the time, Are Swiss fleeing deposits and hoarding cash, pointing out that the demand for Swiss francs is often driven by safe haven concerns. The observed increase in 1000s might therefore have very little to do with the SNB hitting the effective lower bound and everything to do with worries about falling equity prices, China, deteriorating credit quality, and more.

With many of these concerns subsiding in 2016, one might expect the safe haven demand for 1000 franc notes to be falling again. And that's exactly what we see in the chart below; the Swiss are accumulating notes at a decelerating pace, even though the SNB has not relaxed its negative deposit rate one iota.

What these charts all show is that the effective lower bound to central bank deposit rates has not yet been engaged, even after many months in negative territory. You can be sure that the global community of central bankers is watching this data too; it is telling them that, should the need arise, their respective interest rates can be pushed lower than the current low-water mark that has been set by the SNB and Danmarks Nationalbank at -0.75%, say to -0.85% or even -1.0%.

There are a few factors that might be dampening the demand for paper notes. Remember that the Swiss and the Japanese have installed cash escape inhibitors; mechanisms that reduce the incentive for banks to convert central bank deposits into cash. I've written about them here.

Secondly, the SNB and the BoJ, along with the Danes, have set up an array of interest rate tiers. While the marginal deposit earns a negative rate, the majority of the tiers are only lightly penalized or not penalized at all. This tiering represents a central bank subsidy to commercial banks; in turn, banks have passed this subsidy on to their retail depositor base in the form of higher-than-otherwise interest rates, the upshot being that very few banks have set negative deposit rates on retail customers. This has helped stifle any potential run on deposits.

Tiering and cash escape inhibitors have helped dissuade cash withdrawals by two members of the public, retail depositors and banks, but not the third; large non-bank institutions. That these latter institutions haven't bolted into cash shows that the natural costs of storing wads of paper are quite high, and that the effective lower bound quite deep.

Monday, May 16, 2016

Aggressive vs defensive debasement

COINING IN PARIS c. 1500 (From a French print c. 1755)

Not all debasements, or reductions of the precious metals content of coins, are equal. Among scholars of medieval coinage, there is an interesting distinction between aggressive, or bad debasement, and defensive, or good debasement.

Let's take defensive debasements first.

In medieval times, the minting of coins was usually the prerogative of the monarch. Any member of the public could bring their silver bullion or plate to the mint where the monarch's agents would strike a fixed amount of coins from that silver, returning the appropriate number of coins to the owner but taking a small commission for their pains.

A decline in the quality of coin was a fairly natural feature of medieval societies. As silver pennies passed from hand to hand, oil and sweat would remove small flecks of metal.  Compounding this deterioration were less honest methods of removing small bits from each coin, like clipping. Nicholas Mayhew, a numismatist, estimated that each year 0.2% of the coinage's silver content was lost; more colourfully, 'seven tons of silver vanished into thin air' during every decade in the fourteenth century. Less conservative estimates go as high as 1% per year.*

Using Mayhew's 0.2% rate of decline, if the king or queen's mints manufactured pennies that contained 2 grams of silver in 1400, these 1400-vintage pennies might contain just 1.81 grams by 1450, fifty years later.

This created a huge problem. Long before 1450 people would have lost their incentive to bring raw silver to the mint to be made into new coins. Let's say a merchant in 1450 owed 10 pence to his supplier. One option was for the merchant to bring enough silver to the mint to get ten new pennies and then pay off his debt. With the king maintaining his fifty year-old policy of turning two grams of silver into a new penny, that meant the merchant had to bring 20 grams to be minted.

But the merchant had a better alternative; buy ten pennies of the 1400-vintage that together contained just 18.1 grams of silver (1.81 x 21) and then pay off the supplier. The key here is that all pennies, whether they be 1400-vintage or 1450-vintage pennies, passed at face value as legal tender. Thus the merchant's creditor had to accept any penny to settle the debt, bad or good. The merchant's decision was therefore a simple one. Far cheaper to pay off the debt with ten 1400-vintage pennies, the equivalent of 18.1 grams of silver, than ten new pennies, which contained 20 grams.

Because the king's mint was effectively providing too few new pennies for a given quantity of silver, no one would ever bring silver to the mint. (If you work it out, you'll see this is in instance of Gresham's law.) And with no new coins, coin shortages emerged. The legacy coinage had to circulate ever faster to meet society's demand for a medium of exchange, and this would have only increased its rate of depreciation. And a faster decline in the quality of the coinage made them more susceptible to counterfeiters, which only reduced their legitimacy. An inflationary spiral emerged.

The way to simultaneously halt inflation and encourage the creation of new pennies was to introduce a defensive debasement. If, in 1451, the royal family announced a debasement of the coinage so that its mint now struck pennies that contained, say, 1.75 grams of silver rather than 2.0 grams, then people would start bringing silver to the mint again. After all, our merchant could take ten worn-out 1400-vintage pennies that contained 1.81 grams to the mint, have them recoined into ten new pennies with 1.75 grams of silver, pay his debt, and still have some silver left over. The entire generation of old worn out coins would be brought to the mint to be replaced with a new generation of harder-to-counterfeit and clip pennies.

In sum, to ensure a steady supply of new coins the king or queen had to debase the coinage ever few decades. Meir Kohn calls this a ratification of the natural deterioration in the silver content of coins; "defensive debasements did not cause the gradual inflation that took place so much as ratify it." This was sound monetary policy.

Aggressive debasements, one the other hand, were unsound monetary policy.

You may have noticed that by debasing the penny from 2.0 to 1.75 grams, the monarch would be drawing large amounts of silver and old pennies into his or her mint to be turned into new pennies. After all, why would anyone pay debts with pennies that contain 1.81 grams when they can bring them to the mint to be recoined into pennies that contain just 1.75 grams? Another way to think about it is this: if a horse is selling for a pound (where a pound was defined as 240 pennies), why pay for it with 240 pennies minted in 1400 containing a total of 434 grams when they can be first reminted into 240 new pennies that contain just 420 grams, these new pennies being just as acceptable in trade as the old ones? In other words, better to buy a horse for 420 grams of silver than 434.

Debasements were profitable for the monarch. As I mentioned earlier, the royal family levied a fee on all silver brought to his mint. This fee is referred to as seigniorage. In more modern terms, think of a mint as a pipeline where the owner takes a cut on throughput. So by debasing the coinage, the monarch would dramatically increase mint throughput and therefore boost seigniorage revenues.

When a monarch debased the coinage at a much faster rate than the natural rate of wear and tear, he or she wasn't just playing catch up, this was aggressive debasement. One of the most aggressive debasements of all, that of Henry VIII, involved ten debasements between 1542 and 1551, each in the region of 30-40%. These diminutions were so successful in driving silver to the royal mints that Henry had to erect six new mints just to meet demand, according to Kohn.

The motive for aggressive debasements was almost always the funding of wars. As John Munro points out, securing "additional incomes from taxes, aides, loans, or grants from town assemblies, ‘estates’, or other legislative assemblies was difficult and usually involved unwelcome concessions, and this was not necessarily forthcoming." The mints, however, were firmly under the control of the royal family and were therefore a trustworthy form of revenue.

In Henry VIII's case, his debasement revenues were used to fund wars in the 1540s against Scotland and France. But his debasements were bad monetary policy as they caused rampant inflation, specifically a 123% rise in the English consumer price index from 1541 to 1556.


By the way, it was in response to these aggressive debasements that one of the earliest economic tracts, Nicolas Oresme's de Moneta, was written. Oresme's essay, which dates to 1360, was a response to the aggressive monetary policies of Philip VI and John II. According to Robert Mundell, John II debased the coinage 86 times. Wrote Oresme:
I am of opinion that the main and final cause why the prince pretends to the power of altering the coinage is the profit or gain which he can get from it; it would otherwise be vain to make so many and so great changes.
All monetary policy debates since then, including the explosion of words on the econ blogosphere beginning after the 2008 crisis, are versions of the one that Oresme engaged in: what constitutes good debasement and what constitutes bad? While the content of the debate has changed, the structure is pretty much the same.

* I get this from John Munro.
Note: I have an old post from 2013 on defensive debasement. This post is different because it works out the aggressive side of the debasement equation.

Friday, May 6, 2016

What makes medieval money different from modern money?

What's the main difference between our modern monetary system and the system they had in the medieval ages? Most of you will probably answer something along the lines of: we used to be on a commodity standard—silver or gold—but we went off it long ago and are now on a fiat standard.

That's a safe answer. But the fiat/commodity distinction is not the biggest difference between then and now.

The biggest difference is that in the medieval age, base money did not have numbers on it. Specifically, if you look at an old coin you might see a number in the monarch's name (say Henry the VIII) or the date which it was minted, but there are no digits on either the coin's face or obverse side indicating how many pounds or shillings that coin is worth. Without denominations, members of a certain coin type could only be identified by their unique size, metal content, and design, with each type being known in common speech by its nickname, like testoon, penny, crown, guinea, or groat. Odd, right?

By contrast, today we put numbers directly on base money. Take the Harriett Tubman note, for example, which has "$20" printed on it or the Canadian loonie which has "1 dollar" etched on one side.

This seemingly small difference has huge consequences for the monetary system. I'll illustrate this further on in my post by having a modern central bank adopt medieval-style numberless money. The interesting thing is that, contrary to our prejudices about commodity-based money, the medieval system had the potential to be a highly flexible monetary system, far more capable of coping with shocks than our current one, and by implementing medieval money, a modern central banker would get a powerful tool to help in his or her efforts to keep inflation on target.

But first, here are a few more important details about the medieval monetary system. Back then, sticker prices and debts were not expressed in terms of coins (say groats or testoons) but were always advertised in the abstract unit of account, pounds (£), where a pound was divisible into 20 shillings (s) and each shilling into 12 pence (d).  Say that Joe wants to settle a debt with Æthelred for £2 10s (or 2.5 pounds). In our modern monetary system, it would be simple to do this deal. Hand over two coins with "1 pound" inscribed on it and ten coins with "one shilling" on them. Without numbers on coins, however, how would Joe and Æthelred have known how many coins would do the trick?

To solve this problem, Joe and Æthelred would have simply referred to royal proclamation that sets how many coins of each type comprised a pound and a shilling. Say Joe has a handful of groats and testoons. If the king or queen has proclaimed that the official rate is thirty testoons to the pound and eighty groats in a pound, then Joe can settle the £2 10s debt with 60 testoons and 40 groats or any another combination, say 75 testoons. If the monarch were to issue a new proclamation that changes this rating, say a pound now contains forty testoons, then Joe's debt to Æthelred must be settled with 100 testoons, not 75.

To sum up, in the medieval ages the method of determining the content of the unit of account was divorced from the physical objects that were in circulation. Rather than appearing on the coin, the proper ratings were printed up on a royal decree. By contrast, in our modern era monetary authorities have stopped the practice of remotely defining the unit of account in favor of striking it directly onto physical and digital objects.


Let's try to get a better feel for what it would be like as consumers if we didn't have numbers on our modern money. Let's convert today's standard into a medieval standard using U.S. currency as our example. Start by removing the words "one cent" from all one cent coins. Americans take to calling them Lincolns, since Abraham Lincoln is on the obverse side. Next, let's strip all mentions of "one dollar" and "$1" off the dollar bill, which now goes by the moniker "a Washington." Finally, blank out any incident of "100" and "one hundred" from the one hundred dollar bill. Say that people now call it the Franklin. (For simplicity, assume the $10s, $20s etc don't exist.)

Say we go shopping for groceries and get a bill for $302.15. Without numbers on our bills and coins, how are we to know how many Lincolns, Washingtons, and Franklins we should pay?

We would start off by launching our Federal Reserve app to check how many Lincolns, Washingtons, and Franklins the Fed has decided to put in the dollar unit of account that day. Now it could be that the Fed is rating the dollar as 100 Lincolns, 1 Washington, and 0.01 Franklins, which would correspond to the ratios we are so familiar with. In which case, the grocery bill can be discharged with three Franklins, two Washingtons, and fifteen Lincolns. Easy.

But there is no reason that the Fed couldn't be setting a different definition of the dollar unit of account that day. Say that the Fed has re-rated the dollar so that it is now worth 140 Lincolns, 1.4 Washingtons, and 0.014 Franklins. It would be as if an old dollar bill now had $0.714 printed on it instead of $1 (where 1/$1.4=$0.714). In effect, this increases the value of the dollar unit of account in terms of physical cash or, put differently, reduces the purchasing power of the Washington/Lincoln/Franklin. The $302.15 grocery bill must now be discharged with four Franklins and twenty-three Washingtons, where the Franklins cover the first $285.71 and the Washingtons cover the remaining $16.43.

So that's pretty interesting, no? In a modern version of the medieval monetary system, not only do we need to keep track of the coins and notes in our wallet, we also need to follow the Fed's ratings, say with an app. It's cumbersome system but it seems to have worked.


As for the central banker's perspective, let me repeat my earlier point: if the Fed (or any other central bank) were to adopt a medieval system, it would have far more flexibility in hitting its inflation targets than it currently does. Right now, an inflation targeting central bank like the Fed can only target the CPI by modifying the nature or supply of the physical and electronic tokens in circulation, say by altering the quantity of these tokens, changing their interest rate, or shifting their peg (to gold or some other currency). With denominations effectively etched onto coins and printed onto notes, the ability to directly manipulate the unit of account by adjusting the number of bills and coins per dollar has been taken away from it.

By blanking out the numbers and defining the coin/bill content of the dollar remotely, the Fed gets an extra degree of freedom. If it wants to create inflation, it simply posts an alert on its app that the dollar will now contain fewer Lincolns, Washingtons, and Franklins than before. In response, stores will quickly increase their sticker prices so that they receive the same real quantity of payment media as before. Voila, the CPI rises. To create deflation, the Fed does the opposite and puts more notes and coins into each dollar. Just as it currently schedules periodic interest rate announcements, the central bank might issue these edicts every few weeks or so.

In some sense, central banks already engage in alterations of the bill/coin content of the dollar when they redenominate currency. But redenominations are rare, usually occurring during hyperinflations when a central bank cancels all existing currency and issues new bills and coins with a few zeros lopped off. They aren't a regular tool of monetary policy because a continuous series of small redenominations would involve constant recoinages and printing of new notes, an expensive way of doing monetary policy.

By removing numbers from bills/coins so that only blanks circulate and then defining the value of those blanks remotely, redenominations become a cheap tool of monetary policy, much as they were in the medieval days. The Fed wouldn't have to call in all bills and coins and print/mint new ones, it would simply announce a re-rating on its app.


In any case, I hope you can see now that the medieval monetary system was far more complex than we commonly assume it to be. Because we are so used to having number on our bills and coins, we'd be quite lost if we were transported back to the 1500s. As for the monetary authorities, they had an extra set of powers that today's central bankers don't have, the ability to directly define the unit of account. Hypothetically, these powers could have been used to target stable prices or nominal income. In reality, monarchs had different motives, including the funding of wars, so they used the tool for that purpose. But that's another story.

[1] The Spufford Currency Exchange
[2] Meir Kohn: Medieval and Early Modern Coinage and its Problems
[3] John Munro: The Technology and Economics of Coinage Debasements in Medieval and Early Modern Europe