Wednesday, June 13, 2018

Sovereign Money, Narrow Banks, Digital Currency, etc.

The recent Vollgeld (full money) or sovereign money referendum in Switzerland, which was just voted down, is quite interesting, as it raises several fundamental issues in monetary economics and central banking. What's money? Is "money" a useful concept to be bandying about? Who should have the right to issue money? What's a bank? Are banks inherently unstable? What's a central bank for anyway?

I read a few blog pieces and news reports on Vollgeld. Some of the blog pieces, like David Beckworth's, contained essentially nothing I could agree with. Others, like one by Morgan Ricks, were pretty interesting.

What was the Vollgeld proposal? In Switzerland, as you may know, they like referenda. These are conducted for all three levels of government, and there are typically several a year at the federal level. If you're upset about something, and can get enough people to sign a petition, the initiative is sent to the electorate (you won't have to wait for an election), and it's either voted up and made law, or voted down. This specific initiative - Vollgeld - was a proposal that would have banned private banks from the retail payments business. Under Vollgeld, a private bank could still issue various liabilities - long-term debt, short-term paper, savings accounts, for example, but presumably not accounts subject to payment by debit card or check. What was the reasoning? Apparently, we shouldn't trust private banks. They're risky - I assume in both the institutional and systemic sense. And there seems to be some notion that giving a private financial institution the right to issue "money," is giving it monopoly power. Why should we be endowing these elite bankers with special privileges?

But what are people to do if the banks can't provide them with transactions services? The Vollgeld proposal was to transfer that provision to the central bank, which would be then be in the business of setting up accounts, issuing debit cards, making sure accounts aren't overdrawn, filling ATMs with cash, etc. And why not, the Vollgeld proponents might say. After all, those elite bankers have had checking accounts (i.e. reserve accounts) with the Swiss National Bank since that institution was set up. Surely it would be more democratic if reserve accounts were available to everyone.

On the face of it, Vollgeld seems not so different from traditional narrow banking proposals, which have been around for a coon's age. The weird thing is that narrow banking proposals usually came from the right wing, while Vollgeld appears to be a lefty proposal. Milton Friedman, for example, had a narrow banking proposal - the 100% reserve requirement. Friedman thought that controlling the stock of money is everything, and if each unit of transactions deposits were backed by one unit of central bank reserves, the central bank could perfectly control the total stock of money - reserves plus transactions accounts plus central bank-issued currency. So, with Friedman's proposal the private banks could still issue the transactions accounts, not the central bank, but it's effectively the same as a Vollgeld-type proposal. And the rationale isn't so different. The Vollgeld people are concerned that a private institution is issuing what they think is a public good - money - and Friedman agrees, basically (or would if he were still with us).

Narrow Banking in its extreme form has from time to time entered the realm of economic quackery. Murray Rothbard, for example, not only did not trust private banks to create money, he didn't trust the central bank either. Rothbard apparently had some influence on Ron Paul and the End the Fed movement. Basically, Rothbard wanted to re-establish the gold standard, and have a 100% reserve requirement - in gold.

The typical benefits attributed to narrow banking - other than what Friedman had in mind - can be found in some of the concerns of the Vollgeld proponents. Banking is sometimes thought to be inherently unstable - an idea that might occur to you if you were a student of US banking history, for example. A remedy for this problem, instituted first in the United States, is deposit insurance. But, we know that deposit insurance breeds another problem - moral hazard. Banks will take too much risk, if they're not somehow prevented from doing so. So one regulatory intervention breeds others - capital requirements for example. Solution: Adopt narrow banking and transactions deposits are safe - you can do away with deposit insurance.

What's the problem with that? Well, in typical narrow banking proposals, a concern is that the legal restriction on banks makes financial intermediation less efficient. That is, through the wonders of asset transformation, conventional private banks convert illiquid loans and other illiquid assets into liquid transactions deposits. This intermedation process is an efficient way to channel savings into productive investment. If people are forced to hold their transactions balances in accounts backed by central bank or government liabilities, then real interest rates will be higher on loans and other private debt instruments, and there will be less capital accumulation.

The other problem is that narrow banking rules, including Vollgeld, can be skirted by private financial intermediaries. This is a standard game in the financial sector. Some regulator puts restrictions on some type of financial intermediary activity, and some enterprising banker finds a way to design a financial product that performs roughly the same function as what is being regulated, but the new product falls outside of the regulation. In the United States, money market mutual funds came into being because of regulations on transactions deposits at commercial banks. Sweep accounts were devised to get around reserve requirements and the prohibition of the payment of interest on demand deposits. The shadow banking sector grew because costs were lower than in regulated commercial banking.

A curious feature of Switzerland is that, if the Swiss National Bank were to issue transactions accounts, this would actually not be narrow banking in the usual sense. To illustrate why, consider Canada (if you dare). In Canada, for example, most of the assets held by the Bank of Canada are obligations of the government of Canada, and the balance sheet is small, at 6.5% of GDP. So, if the Bank of Canada issued transactions accounts to consumers, that would be a narrow banking activity, provided nothing changed in the composition of the Bank's asset portfolio. The Swiss National Bank, by contrast, holds a portfolio that includes a large amount of equity, and the majority of the portfolio is not denominated in Swiss currency. The Swiss National Bank holds the sovereign debt of other countries in Europe, North America, and Asia, and it holds private debt. Further, the balance sheet of the Swiss National Bank is very large, at 117% of Swiss GDP at the end of 2017.

So, if Vollgeld had passed, what the SNB would be doing would not be narrow banking as traditionally envisioned. In fact, the Swiss National Bank, given its large and unusual asset portfolio, wouldn't look like any bank I know of under Vollgeld. It would have been some sort of stock mutual fund/foreign currency-denominated bond fund offering transactions accounts. Though the Swiss National Bank is currently well-capitalized, I doubt if it would pass the regulatory scrutiny of any bank regulator. In any case, the Swiss National Bank campaigned actively against Vollgeld.

Another curious aspect of Vollgeld was that the movement was driven in part by the idea that banking should be democratized. Somehow private banks were thought to be falling short in the public service department. But the residents of Switzerland already have the option of holding transactions accounts - and savings accounts too - in a government financial institution. If you've been to Switzerland, you should know what this is - it's the Swiss post office, Swiss Post. From their web site, it appears that Swiss Post also offers financial services related to pensions and real estate. Wikipedia says Swiss Post is the country's second-largest employer, and Post Finance, its financial arm, is the fifth largest retail financial institution in Switzerland. So, given that there is already a government-owned option in retail payments services, why would the Vollgeld people want to take the extra step of banning private retail payments outright? Beats me.

For all central banks in the world, changing technology, and evolution in retail payments has become an important issue - if not the key issue for the future of central banking. Central-bank-issued currency is losing ground to debit and credit cards in retail payments, in spite of the fact that currency outstanding has in some instances increased. Here's the currency/GDP ratio for the US:
So, that bottomed out at about 4.1% in the 1980s, and currently stands at 8.1%. But, about 80% of that is in $100 denominations. As people like Ken Rogoff like to point out, the primary liability of most central banks - what finances the asset portfolio of a small-balance-sheet central bank like the Bank of Canada - is being used as a means of payment and store of wealth by criminals. Getting rid of large denominations, or withdrawing all currency from circulation, would increase the cost of committing crime, and might make us all better off (except the criminals of course).

But, currency is still used intensively, in legitimate ways, by a significant fraction of the population, who tend to reside in the bottom tail of the income distribution. And currency can be useful when everything else fails - computer networks go down, or the power fails altogether. We don't want to shortchange the poor (literally), or shoot ourselves in our collective foot. A possibility is that central banks could issue some form of digital currency, which is what most people are considering, I think, when they discuss "sovereign money," not the more draconian Vollgeld approach.

So, if central banks were to issue a new type of liability, what form could that take? Well, it could be a kind of cryptocurrency. As with Bitcoin, for example, central banks could issue a liability that could be exchanged in decentralized fashion, on a distributed ledger. While central banks may be able to correct some of the potential stumbling blocks associated with cryptocurrencies, what we know about blockchain technology currently suggests that it can't handle a large volume of transactions, and that it is far too costly - in pure energy costs - to provide the correct incentives. So, a central bank digital currency would have to be centralized - a stored-value-card technology for example. But then the central bank would have to provide an interface whereby value could be downloaded and uploaded at decentralized locations. Possibly this could be done through the private banking system, and possibly central bank digital currency could compete with privately-issued digital currencies.

But why should the central bank be in the business of retail payments at all? Why not leave that entirely to the private banking system? Indeed, there were historical instances of stable and successful private money systems. For example, before 1935 in Canada, retail means of payment consisted of small denomination government of Canada paper notes and coins, larger denomination paper notes issued by private chartered banks, and transactions deposits at those same chartered banks. If currency is an anachronism, we could do away with it, and leave the retail payments market to private financial institutions, perhaps with some constraints requiring that basic services be provided to low-income people.

What would monetary policy be in a world in which the central bank doesn't issue liabilities that Milton Friedman would recognize as "money?" Here, it's helpful to get away from the idea that there is some stuff we should call "money," and that such stuff is produced through magic - creation from "thin air," apparently. Those ideas - propagated sometimes by otherwise-respectable economists - aren't helping us much, and are seriously misleading. In reality, assets exist on a multidimensional spectrum, and the central bank's ability to execute what we think of as monetary policy - inflation control, crisis intervention, smoothing of financial conditions to enhance real economic activity - could be unimpeded even if the central bank does not issue a retail payments instrument.

Central banking works because the central bank is transforming assets - its a bank after all, which has assets and liabilities that differ according to maturity and liquidity. As long as the asset transformation is something that can't somehow be undone by private actions, it works. Governments issue interest-bearing debt for which close private sector substitutes don't seem to exist. For example, in the US, the repo market loves Treasury securities. So, we might imagine central banks issuing an array of interest bearing obligations - from overnight assets like the reverse repos issued by the Fed, to three month bills, to 10-year bonds - to finance a portfolio of government debt. In principle, those central banks should be able to target short-term market interest rates, just as they do now.

I've touched on a host of questions that central banks need to know the answers to, and there is some urgency in providing answers. Unfortunately, we can't get the answers by running a regression, as these are fundamental questions having to do with institutional design and technologies that we have little or no experience with. This is where theory comes in, and part of why it's useful - sometimes the only game in town. Get out your models, monetary theorists!


  1. This is the good old boundary problem by Charles Goodhart at work:
    «Some regulator puts restrictions on some type of financial intermediary activity, and some enterprising banker finds a way to design a financial product that performs roughly the same function as what is being regulated, but the new product falls outside of the regulation.»

    As a result, John NugĂ©e has thought about reconsidering basic concepts instead of issuing rules after rules…

    In the same spirit, Jonathan McMillan thought that Positive Money (another narrow banking strand) is asking for too little in the end:

  2. 1. The Pre-1935 Canadian banking system had multiple liability. Of course, post-1935 Canadian banks don't have that restriction, and they have been remarkably stable.

    2. The McMillan proposal is a new one to me. Didn't know there were people who wanted to get rid of banking altogether. How could that go wrong?

  3. "... in the US, the repo market loves Treasury securities. So, we might imagine central banks issuing an array of interest bearing obligations - from overnight assets like the reverse repos issued by the Fed, to three month bills, to 10-year bonds - to finance a portfolio of government debt. In principle, those central banks should be able to target short-term market interest rates, just as they do now."

    This is how banks get into trouble. The federal reserve bank system is a network of 12 private banks that issue currency, take deposits from their member institutions (national chartered commercial banks, and state chartered commercial banks that have elected to become members), pay dividends at a rate of 6% on permanent capital to the members, set interest rates and engage in money-market operations. If the federal reserve banks were to issue debt securities in their own name in order to fund government operations, i.e., to fund primary deficits, what would prevent the federal reserve banks in connivance with politicians from engaging in a never ending yet legal Ponzi scheme relating to the federal government debt? But isn't this the situation we have today, with the notable exception that the federal reserve banks do not yet issue long-term debt in order to fund the government's primary deficits?

    John Cochrane, in a sequence of learned papers, places at the core of his theory of interest rates and inflation the relationship between the government's current outstanding debt and the value of that debt as a function of the current period expectation of the discounted time sequence of future government primary surpluses over time stretching from today to infinity. The value of the debt securities (Treasury notes and 'bonds'), in his view, is predicated on the notional existence of future government primary surpluses of sufficient quantity and duration to give meaningful substance to the debentures the government has issued.

    Now, let us suppose that there is no meaningful likelihood that future governments will produce those primary surpluses, and, further, let us suppose that this becomes broadly known to investors. Then applying John Cochrane's valuation formula, the current value of the government debt obligations is nil (zero). But how could this happen, if the government's banker is able to issue currency notes in fulfillment of the government's debt obligation? By printing paper money, the government fulfills its obligations and passes onto to its citizens the cost of those primary deficits in the form of depreciation of the purchasing power of their currency and credit deposit holdings. In nominal currency units, the value of real assets increases under such a regime. We describe this as 'inflation'.

    Adam Smith, in his Wealth of Nations, referred to it as the 'corn price of gold' in connexion with the money inflation in Spain and Portugal that resulted from the transfer of commodity silver and gold that flooded into the Iberian Peninsula from the Americas in his time. The principle is the same. A 2% inflation target is akin to monetizing the federal debt and facilitating a political Ponzi scheme whereby the government is relieved of the obligation to produce primary surpluses of sufficient quantity and duration to retire those debts.

    [... cont'd.]

  4. [continuation ...]
    One can, as John Cochrane has done, construct elaborate hypotheses around the relationship between nominal interest rates, real interest rates, and expected inflation rates, but the foundation is the concept that the government debt obligations outstanding are supported by expectation that the government will eventually retire that debt by producing future primary surpluses of sufficient quantity and duration in the future. If, instead, the government has no intention or capacity to do so then the models collapse and the hypothesized relationships are rendered meaningless.

    What then is money? If government bonds are worthless, as they would be under such a scenario, then it must be the case that paper money and coin would be equally worthless for they are naught but government obligations in a different form. Thus, the essential question comes down to the intentions of government with respect to its capacity to produce the primary surpluses necessary to support the value of its debt obligations. No amount of mathematical construction or theorizing, however elegant, will answer that which is at its core a political problem of considerable import. To John Cochrane’s credit, he, along with some others, realizes this.

  5. “…Adopt narrow banking and transactions deposits are safe - you can do away with deposit insurance. What's the problem with that? ... If people are forced to hold their transactions balances in accounts backed by central bank or government liabilities, then real interest rates will be higher on loans and other private debt instruments..."

    Very much agree. You may be interested in this old (but sitll valid) work: