Wednesday, June 27, 2018

Failed monetary technology

Archaic and ignored monetary technologies can be very interesting, especially when they teach us about newer attempts to update our monetary system. I recently stumbled on a neat monetary innovation from the bimetallic debate of the late 1800s, Nicholas Veeder's Republic of Eutopia coin:
If you've read this blog for a while, you'll know that I like to talk about monetary technology. Unlike financial technology, monetary tech involves a technological or sociological upgrade to the monetary system itself. And since we are all unavoidably users of the monetary system—we all think and calculate in terms of our nations unit of account—each of us is immediately affected by the change.

Veeder's Eutopia coin is an old monetary technology that was never adopted. More recent examples of unadopted (or as-yet not adopted) montech include Fedcoin, NGDP futures targeting, or Miles Kimball's technique for evading the zero-lower bound, which would decouple the value of paper money from electronic money. Examples of recent monetary tech that went on to be adopted include the switch from paper to plastic banknotes, the replacement of older end-of-day clearing systems to real time gross settlement systems, and inflation targeting.

Fintech is more limited in scope than monetary tech. Only that portion of the population that uses these innovations is affected—everyone else's financial habits continues on as before. Recent examples include bitcoin, p2p lending, and roboadvisors. (If bitcoin ever became the standard unit of account, it would have made the trek over to becoming monetary technology, and not just fintech.)


To make sense of Veeder's Republic of Eutopia coin, we need to understand the problem that his monetary innovation was meant to solve. Most nations were on a gold standard by the 1870s, and with the price of gold rising, the world price level was generally falling. This development provided an unexpected boost to the creditor class, who were owed gold, while hurting the debtor class, who owed gold. A higher price for the yellow metal meant that the loan contract to which a debtor had signed their name now required them to work that much harder to pay it off.

In that context, a broad popular movement for the remonetization of silver emerged. Prior to being on gold standards, nations were generally on a pure silver standard or a bimetallic standard. On a gold standard the debtor class had only one way to settle the debt, by providing the proper amount of gold coins. But if silver coinage was reintroduced at the old rate of sixteen-to-one, debtors could instead sell their labour to buy cheap silver, have it minted into legal tender silver coins, and use those silver coins to pay off the debt. Paying their debts with silver rather than gold meant they'd have a bigger amount of wealth remaining in their pocket.

The movement to restore bimetallism wasn't purely a populist one. The smartest economists of the time--folks like Irving Fisher, Leon Walras, and Alfred Marshall--also preferred bimetallism. A bimetallic standard recruits more monetary material into service than a gold standard. This is advantageous because, as Fisher put it, it "spreads the effect of any single fluctuation over the combined gold and silver markets." In other words, the evolution of the price level under a bimetallic system should be more stable—and thus more fair—than under a monometallic system, since it can absorb larger shocks.

The problem with bimetallism is that it very quickly runs smack into Gresham's law. The traditional way to bring the two metals into service as monetary material was to offer to mint both high denomination gold coins and lower denomination silver coins. So if a merchant needed £20 worth of coins, he could bring either a chunk of raw gold to the mint, or an even bigger chunk of pure silver, and the mint would convert either chunk into £20 for him. The specified amounts of raw silver or raw gold that were required to get a certain number of £-denominated coins constituted the mint's official gold-to-silver exchange rate.

Inevitably the market's gold-to-silver exchange rate would diverge from the mint's official exchange rate, effectively over- or undervaluing one of the two metals. In this situation, no one would bring any of the overvalued metal to the mint to be turned into coins. After all, why bother minting a chunk of gold (assuming the yellow metal was the overvalued one) into £20 worth of coins if that same amount of gold has far more purchasing power overseas? The overvalued metal would thus disappear as it was hoarded and exported, leaving only the undervalued metal in circulation. A monometallic standard had accidentally emerged, and all the benefits of bimetallism were for not.

To prevent Gresham's law from being engaged, the mint had to constantly adjust its official rate so that it stayed in-line with the ever-evolving market rate. Not only would these changes have been politically costly, but they would required an expensive series of recoinages in order to ensure that coins always had the proper amount of silver or gold in them.


Enter Veeder's Eutopia coin. Nicholas Veeder was no economist, but an executive at C.G. Hussey, a copper rolling mill in Pittsburgh. In 1885, he published a pamphlet with the wordy title Cometallism: A Plan for Combining Gold and Silver in Coinage, for Uniting and Blending their Values in Paper Money and For Establishing a Composite Single Standard Dollar of Account.

Rather than defining a dollar as simultaneously a fixed amount of gold OR a fixed amount of silver, Veeder's pamphlet suggested defining it as a fusion of the two together. Specifically, Veeder's dollar was to contain 12.9 grains of gold AND 206.25 grains of silver. It's worth noting that under a proposed cometallic standard, paper dollars needn't be redeemed with actual Eutopia coins, but could be converted into separate silver and gold bars or coins. The important rule was that each dollar's worth of debt had to be discharged with 12.9 grains of gold and 206.25 grain of silver.

A model of a cometallic gold certificate, from page 60 of Veeder's pamphlet on cometallism

Veeder's cometallic scheme was a neat way to keep all the benefits of bimetallism with none of its drawbacks. Cometallism would draw on the combined supplies of the gold and silver markets, so that the system would be much more elastic than a pure gold standard, and thus fairer to both creditors and debtors. At the same time, Gresham's law would be avoided. Under traditional bimetallic coin systems, the mint established an exchange rate between the two metals. This rate inevitably became the system's undoing when it diverged from the true rate.

But a mint that was operating under a cometallic standard would only accept fixed quantities of silver AND gold before it would mint a $1 coin, and so it would no longer be setting an exchange rate between the two precious metals. The undervaluation of one of the metals, a key ingredient for Gresham's law, could never emerge under cometallism.


A year after Veeder published his pamphlet, Alfred Marshall—one of the world's leading economists—described a remarkably similar system. Here is part of his response to the Royal Commission on the Depression in Trade and Industry in 1886, which had been convened to address the Long Depression:
"I propose that currency should be exchangeable at the Mint or Issue Department not for gold, but for gold and silver, at the rate of not £1 for 113 grains of gold, but £1 for 56^ grains of gold, together with, say, twenty times as many grains of silver. I would make up the gold and silver bars in gramme weights, so as to be useful for international trade. A gold bar of 100 grammes, together with a silver bar, say, twenty * times as heavy, would be exchangeable at the Issue Department for an amount of the currency which would be calcalated and fixed once for all when the scheme was introduced. (It would be about .€28 or .€30 according to the basis of calculation)."
Marshall's proposal was later dubbed symmetallism. (I wrote about it here.) If you study monetary systems, you'll run into the gold & silver basket idea sooner or later. The concept is invariably refereed to as symmetallism (and not cometallism) and attributed to Marshall (not Veeder). In the 1800s academics were not required to provide references, and from what I understand plagiarism was rampant. Did Marshall develop his idea separately from Veeder, or did he rip it off? Whatever the case, Veeder was an unknown executive at a small manufacturing concern, whereas Marshall a world famous academic. Celebrity carried the day.


Interestingly, Veeder himself probably borrowed the idea, or at least part of it, from someone else. Almost a decade earlier, William Wheeler Hubbell had tried to get the U.S. congress to adopt the so-called "goloid dollar," a coin containing silver and gold alloyed together.
Hubbell owned the patent to the goloid alloy, so he would have made a good profit if the goloid dollar had been adopted by the U.S. Treasury. Unlike Veeder, Hubbell doesn't seem to have been a very good monetary economist, and the case he makes for goloid misses much of nuances of the benefits of bimetallism and the hazards of Gresham's law. He lists a number of advantages for his proposed coin, including: superior durability to gold and silver coins; not susceptible to oxidization (unlike silver); a goloid dollar was smaller than a silver dollar and thus more convenient for consumers to carry around; the mint would be able to make more goloid dollars than silver dollars with its existing capacity; and goloid coins could not be easily melted down for usage in the arts as was the case with gold and silver coins.

Hubbell's idea foundered on the fact that a goloid coin, despite containing gold, has almost the exact same colour as a silver coin. Hubbell's critics believed this set the coin up to be widely counterfeited. A counterfeiter could make a replica with lower gold content, this alteration unlikely to be noticed by the public since the colour of a genuine goloid coin and the fake would be the same.

The difficulties that Hubbell experienced alloying gold and silver were not lost on Veeder. In has pamphlet he mentions that "my first approach, as with many other persons, was to combine the two metals as an alloy for coinage, but, owing to certain difficulties... this idea was soon considered impracticable and abandoned." To avoid Hubbell's color problem, Veeder ended up mechanically wedding the two metals rather than chemically combining them, the Eutopia coin being comprised of a ring of silver and a gold plug embedded inside.


The topic of goloid and Eutopia dollars seem a bit obscure, but the issues of stability and fairness that concerned monetary technologists in the late 1800s remain relevant today.

Today, most western central banks define the national currency in terms of a basket of consumer goods and services rather than a fixed amount of gold (gold monometallism) or a basket of gold & silver (cometallism, symmetallism). This makes a lot of sense. If we want to create a stable monetary standard, one that provides creditors and debtors with an even playing field, better to use a broad basket of stuff that regular people buy than a narrow basket of metals. That way all parties to a contract know many years ahead of time exactly how much consumer goods they will get (if they are creditors) or give up (if they are debtors). Knowing how much gold and silver baskets they will owe or be owed is less relevant to the average person, since gold and silver are a very small part of most people's day-to-day consumption profiles.

There is an important debate going on today about whether to continue defining national currencies in terms of a consumer goods & services basket, or whether to move to something more fluid like a nominal gross domestic product (NGDP), or output. One problem with using a consumer goods basket is that, in the event of a large economic shock that leads to significant loss of jobs, debtors take on all the macroeconomic risk. After all, they owe just as many CPI baskets as before, but have less capacity to meet that obligation because they might not have a job. This doesn't seem like a fair splitting up of risks and rewards.

The nice thing about defining the national currency in terms of NGDP, or output, is that the risk of a large shock, and the associated loss of jobs, is shared between creditors and debtors. This is because if a recession occurs, debtors will owe a smaller amount of real wealth to creditors than they otherwise would. And during a boom, when the job offers are rolling in, creditors will owe more.

Cometallism was never adopted. Perhaps it was a bit too fancy. NGDP is a bit exotic too, but then again so were many forms of monetary technology, until they were actually adopted and became part of the background. We'll have to see what happens.


  1. Good clear post.

    A very early precursor is coins made of electrum (a naturally occurring alloy of gold and silver).

    1. Thanks Nick. Yep, electrum is an even earlier example.

  2. "defining national currencies in terms of a consumer goods & services basket"

    Hmmm. How would this work when one country is rich in food and another rich in labor that needs food to survive?

    One solution is that the labor rich country produces and exports durable goods in exchange for consumable food products. But, how would a commonly valued commodity basket be established?

    Ultimately, in this example, labor time would be traded for labor-plus-a-highly-specialized-resource-availability. It seems to me that human judgement would have a huge role in establishing any possible link between representative commodity baskets.

    It seems to me that local currencies are created when isolation from other economies is desired.

    1. How would it work? Basically, it doesn't; exchange rates float. Human judgment in the form of international commodity markets establishes the trajectories of each basket, but those trajectories differ.

      If the conclusion is that the two countries should form a currency union, I agree. Unfortunately, monetarist economists build their careers studying (and for a few, managing) floating currencies.

    2. Yep, I think Basil is right. Even if one country is "rich in food and another rich in labor", the value of each currency is linked to its own domestic consumption basket, and the two currencies float against each other. No reason this can't work.

  3. Gresham's law and the impossible "trinity" are the same class of error: setting multiple values for a currency. You cannot put your single egg simultaneously into multiple baskets.

    I wonder what the "limit of increasing baskets" is. Monometallism < symmetallism < forex basket < ??
    CPI basket < GDP deflator < ??
    I think you should try to put your single egg into the largest possible basket for at least two reasons. (1) The largest basket is the most stable one. (2) Say's law. Single shipwrecked guy cannot have surplus. Barter economy can have matching surplus and shortage, but no general glut. Monetary economy can have surplus/shortage of money and matching shortage/glut of everything else. If you manage to define money as the right "everything else basket", this can no longer happen.

    1. "Gresham's law and the impossible "trinity" are the same class of error: setting multiple values for a currency. "

      They do sort of seem the same.

      "I think you should try to put your single egg into the largest possible basket for at least two reasons. (1) The largest basket is the most stable one. (2) Say's law. Single shipwrecked guy cannot have surplus. Barter economy can have matching surplus and shortage, but no general glut. Monetary economy can have surplus/shortage of money and matching shortage/glut of everything else. If you manage to define money as the right "everything else basket", this can no longer happen."

      I agree about larger baskets. Although at some point the case can be made for not just broadening the amount of goods & services included in the basket that defines the currency, but to switching to an entirely different standard, say to defining it in terms of NGDP or wage rates.

    2. I think anything to do with GDP is an attempt at creating the "infinite basket". My guess is that targeting the GDP deflator would be a pretty good way to do it.

      NGDP targeting has a few problems. The most obvious is that NGDP is an extensive property; something like NGDP per capita is a better candidate for being targeted.

      A lesser problem is that NGDP targeting fans keep trotting out the idea that the whole thing should be automated via prediction market. That can be hijacked really, really easily---cheaply enough that it could be profitable.

      Pull a number out of thin air, and say that the Fed throws $100 million/year into the prediction market. They divide this among NGDP futures, or options, or whatever, according to some known rule, proportionately with the accuracy of prediction. Thus the market will have capitalization of the level that makes these securities have market-rate yield.

      Now, imagine I'm an economic advisor to the USSR or whatever baddie. I convince the Secretary General to give me a budget of $50 million/year, and I promise to divide this between the "we need to tighten" contracts of the Fed's prediction market. To the extent my promise is credible, arbitrageurs go into the market, and increase its capitalization until yields go back down to market levels. This unexpected overcapitalization is a dead giveaway that the market is being hijacked, but doesn't reveal the direction. However, now 33% of the payout comes from me; thus the market price of the two opposed contracts will shift from the expected 50-50% Fed payout point, to the point where the Fed's rule pays out 25-75%. Depending on how the payout rule works, this can be an enormous shift in the stance of monetary policy. Mwahahaha!

      Ahem. If I can up the stakes to match (or exceed) the full $100 million/year funding stream from the Fed, I can shift the equilibrating point to the 0-100% payout point of the rule, or ... well, to anywhere I want to.

      Now, even if I don't want to break things for fun, I can still break things for profit. I simply build a position (short and long both work) of a few billion dollars and shove monetary policy in the appropriate direction.

      Even if you change the numbers, this failure mode only becomes more expensive, but doesn't go away. And at some point, you start to wonder whether throwing that much money at the prediction market is the right idea, or if you should just hire a few analysts instead and forget the whole thing.

    3. I'm aware of NGDP futures targeting, but don't know enough to defend it or criticize it.

      I do know that Scott Sumner has defended it against the charge of manipulation, for instance here:

      "Another concern is that special-interest groups might try to manipulate the market for financial gain. Thus, a firm that would benefit from faster NGDP growth might sell a large number of NGDP futures contracts short, pushing the Fed to adopt a more expansionary monetary policy. Evidence from field experiments by Robin Hanson and by Hanson, Ryan Oprea, and David Porter suggests that it is difficult to effectively manipulate a prediction market.34 Any attempt at market manipulation opens up profit opportunities to other traders, who would take advantage of a gap
      between the current market price of NGDP futures and the expected future price of NGDP futures.

      "Nonetheless, given the importance of monetary policy, the central bank might want to take extra precautions against market manipulation. One idea would be to limit the net long or short position for any single trader. It would be preferable to allow unlimited trades but then have the Fed take an opposing position for any “suspect” trades. Over time, this system would provide information about whether the Fed’s suspicions of market manipulation were correct. If the Fed tended to lose money on these trades, it would suggest that market manipulation was not the motivation for private traders taking large long or short positions in NGDP futures."

    4. "Any attempt at market manipulation opens up profit opportunities to other traders, who would take advantage of a gap between the current market price of NGDP futures and the expected future price of NGDP futures."

      This is correct iff "manipulation" means trading. My way of manipulation doesn't involve trading in NGDP futures; even if the goal is profit, the manipulator has a position in dollar-denominated bonds (including cash), not futures.

      Think about it as subverting M̶i̶l̶t̶o̶n̶ ̶F̶r̶i̶e̶d̶m̶a̶n̶'̶s̶ Scott Sumner's thermostat. You definitely don't do it by trying to heat or cool the room directly, because the thermostat compensates for that. What you do instead is that you put a chilled icepack on the thermostat, or you put a lit candle under it. Then the thermostat measures a different temperature from its setting, and turns on the heating/cooling to adjust its *measured* temperature back to its setpoint -- but the room will be much warmer/cooler. (Note the direction of the effect; the hot candle cools the room.)

      Concrete contract design (p. 11):
      "The Fed would offer to buy or sell unlimited quantities of NGDP futures at a price equal to one plus the expected GDP growth rate, or $1.0365. When the contracts mature a year from today, their value will equal the ratio of next year’s NGDP to current NGDP."

      That's a complete mess, if you try to figure out its real return. NGDP(t+1)/NGDP(t)*$(t+1)1.00 payout at t+1, for $(t)1+E price at t (E is the target NGDP growth rate). Divide by the latter, and the yield you get is:
      minus real interest rate. Now, look at that NGDP*$ thing, that's RGDP.
      minus real interest rate. Plus liquidity yield of NGDP futures.

      So, a few consequences:
      1: If RGDP expectations grow, people will buy NGDP futures and cause some deflation. This is the effect Scott Sumner is aiming for.
      2: While real interest rates are closely related to real growth, there's some wiggle room. If real interest rates go down (there aren't enough good ways to invest capital), but RGDP growth is stable, then people will buy NGDP contracts and cause a deflation. This is similar to the Great Depression's problem with flight to gold.
      3: If the liquidity yield of NGDP futures improve, people will buy more and cause a deflation. This could be self-reinforcing to some degree, as less money improves the liquidity yield of NGDP futures. Buying $1 NGDP withdraws $1000 in reserves and introduces $1000 in T-bills(?). Is that sufficient to increase demand for NGDP futures by $1? Hopefully not. Scott Sumner mentions risk premia as a similar problem of unusual yield types.
      4: If I promise to buy NGDP futures after expiration and payout (at a time they would be worth $0.00 normally), then people will buy more and cause a deflation. To call back to the first quote: I change the expected future price of NGDP contracts directly, and the current market price follows that.

      P.20-21 describes the "market deepening" I assumed in my previous post. Again, imagine that I (1) add further money to the subsidy pool and thus deepen the market; (2) I change my mind and divide my added subsidy by different rules. This still deepens the market, but also distorts it, if the rule I use to distribute the subsidy isn't exactly symmetric.

  4. Wouldn’t any profits from arbitrage be swallowed by shipping costs? What about theft?

    In the end money should encapsulate the value of an economy’s goods and services which is why commodity money was ineffective. As the economy grew and expanded at an exponential rate, commodity money couldn’t keep up. Creating a coin from gold and silver solved an irrelevant problem, the creation of either more silver coins or more gold coins. But what was essential for the economy was a token that could be stated to have value to reflect the value being created. You could have achieved the same end by mixing silver and copper. The problem with the gold/silver coin was with trying to create a unit that both appreciated significantly in value *and* was common enough to use as a mass currency. That can’t be done without subjecting an economy to regular economic shocks. All roads lead to a common money that has >0 value that expands and contracts to reflect economic activity. The math was always against commodity money, no matter how it was as configured, particularly gold. The social dynamic was always against coins because paper (and plastic) are far more convenient and easy to use. Money is ultimately simply a resource allocation mechanism. The arguments regarding arbitrage are simply distractions IMO.

    1. "Wouldn’t any profits from arbitrage be swallowed by shipping costs? What about theft?"

      I don't follow. Profits from arbitrage? Are you talking about the profits from exporting the undervalued metal overseas?

  5. Yes. Stands to reason that shipping precious metals overseas to exploit arbitrage opportunities would have been very expensive, particularly back then.

    1. Not really. The American colonies used to export grain to Europe even in the 1700s. Compared to that, the shipping fee on literally a ton of "nondescript cargo, definitely not gold!" must have been a negligible fraction of the value of that ton of gold. Alternatively, put a few pounds of gold into your handbag, and compare that to the ticket price.

      The point of an international gold standard was that (1) gold was worth the same everywhere, due to its ridiculously high value density; and (2) if each nation (or even each bank) kept up its currency board to gold, then (1&2) the exchange rate between any two currencies was fixed for practical purposes. It could have been a cowrie shell standard just as well, but the network effect got started around gold, so gold it was. Let's hold a Keynesian beauty contest for the commodity to base our currencies on!

    2. "Stands to reason that shipping precious metals overseas to exploit arbitrage opportunities would have been very expensive, particularly back then."

      From Britain to France? Probably not that expensive. But in general, you are right. Prices would have to move by at least +/- x% before Gresham's law is engaged. Within that band, a bimetallic standard still functions. We have some good examples of bimetallic systems succumbing to Gresham's law (Britain in the 1700s), so we know that the band wasn't terribly wide, even back then.

  6. Let us go through a list of failures.

    The gold standard failed.
    Blue and green stamps were replaced by groupon.
    The first and second bank of the USA failed.
    XRP, the Ripple coin serms to be failing against Bitcoin.
    Judging by the size of the Fed balance sheet, central banking may be failing.
    The greek central bank seems to be failing, and Italy is not looking to hot. On the subject, Illinois failed.
    Endless monetary technology failures, like any other business.
    But Fintech is a wild success, investment money pouring in and fiat banks happy as can be. Fiat bankers sell digital liquidity a monopoly prices to Fintech, nothing unsuccessful about that.