Sumner on a Pure Credit Economy

Scott Sumner has responded to the "hypothesis" of a pure credit money.

Sumner correctly points out that the U.S. certainly does not seem to be in a situation where the demand for base money--currency and reserve balances at the Fed--is withering away. The quantity base money is remarkably high by historic standards. Further, the standard Market Monetarist account of the status quo is that the demand to hold those balances is higher still.

Sumner also correctly points out that if a monetary authority chooses to target interest rates, or anything else, the quantity of money becomes endogenous. It is adjusting the monetary base to hit its target. To respond to a claim that the monetary authority should target base money with detailed explanations that amount to saying that it doesn't, is wrongheaded.

If we did live in a world where the demand for hand-to-hand currency was decreasing and new methods of settling inter-bank clearings were allowing for a reduction in the demand for reserve balances, then this would make holding the quantity of base money constant inflationary. From a Market Monetarist perspective, the necessary course of action would be to decrease the quantity of base money in step with the demand to hold it.

Of course, Market Monetarists have never proposed fixing the quantity of base money and having the price level adjust so that the real quantity adjusts to the real demand. Instead, they favor a target for the growth path of nominal GDP. In this scenario, the nominal quantity of base money would be reduced in step with its falling demand so that nominal GDP would remain on the targeted growth path.

Only when the demand for base money, both hand-to-hand currency and reserve balances, is zero, would controlling the quantity of base money so that it equals the amount demanded be a bit paradoxical. It is in this scenario where Woodford argued that interest rates could be controlled by the relatively conventional channel method. The central bank sets an interest rate at which it will lend to banks by creating balances in their reserve deposits. And it also chooses a lower interest rate that it pays banks for holding those balances. By raising or lowering both, the central bank controls short term interest rates throughout the economy.

Of course, simply paying interest on reserve balances would slow any process that was leading banks to reduce their demand to use them for clearing purposes. Still, if the demand for base money does fall to zero in equilibrium, that doesn't prevent the monetary authority from using open market operations to control nominal GDP, the price level, or some conglomeration of privately-issued deposits that it might call "money."

If the monetary authority wanted to increase inflation, raise nominal GDP, or increase the nominal quantity of some set of monetary assets, it would makes an open market purchase. The quantity of base money rises above zero. The seller's bank would have excess reserves--the desired quantity being zero by assumption. A bank with excess reserves gets rid of them in the usual way. It makes loans or purchases some other asset, or perhaps pays lower interest on deposits. Of course, this simply shifts the reserve balance to some other bank. The "hot potato" continues until whatever the desired nominal target is reached. Notice, that as the banks make loans or purchase securities they are increasing the amount of deposits in the hands of households and firms and as they lower the interest rates on those deposits they are reducing the demand to hold them. The private banking system is creating an excess supply of money.

What is unusual, is that if the demand for base money is zero in equilibrium, the monetary authority must make an open market sale once its nominal target is reached. This reduces the quantity of base money back to zero, the amount demanded in equilibrium.

The opposite scenario is even more peculiar. With the demand for base money being zero in equilibrium, the monetary authority has no assets to sell. How can it make an open market sale? It must borrow a security and then sell it. This is what is called short selling a security. (I am not sure I want to say that base money is negative.)

The buyer's bank then has a reserve deficiency. If the penalty for such a penalty is mild, for example, some penalty interest rate, then the bank with debit balance contracts credit by restricting new loans or selling off securities. Or it could pay higher interest rates on deposits. That just shifts the deficiency to some other bank, and the contraction continues until the monetary authority reaches its target. Again, notice that the banks are contracting the quantity of checkable deposits or raising the interest rates paid on them. The private banks are creating an excess demand for money. Again, when the monetary authority reaches its target, it must make an open market purchase, returning base money back to zero and paying back the security it borrowed.

If the penalty for a reserve deficiency was sufficiently harsh, then the demand for base money would have never fallen to zero. Allowing banks overdrafts at market rates leads to banks desiring to hold no reserve balances in equilibrium. And while the penalty rate suggested above is very similar to a central bank's lending rate in a corridor system, it does not have to be a target. It can be set at a rate higher than some market determined rate. Similarly, if the monetary authority pays interest on reserve balances, that interest rate can float as well, being set below some money market rate.

If the demand for base money falls to zero, a central bank can control interest rates. However, as Sumner explains, it is never sensible to have have an interest rate goal. This leaves the price level indeterminate. The interest rate must be adjusted to control some other nominal quantity. Woodford mostly focuses on an inflation target.

And if the demand for base money falls to zero, treating the growth rate or path of the monetary base as a target would be absurd. Still, there is no need to control interest rates. All interest rates can be left to market forces, and open market operations can be used to control some other nominal quantity--like inflation or the growth path of nominal GDP.

While Sumner's criticism of Ashwin Parameswaran's post were mostly on the mark, he goes too far when he says:

There will probably always be money; a pure credit economy is unthinkable. Without money there is no price level, because the price level is defined as the average price of goods in terms of money.

A pure credit economy has money. It's defining characteristic is that all money is a liability of some issuer. It represents a debt of someone. Checkable deposits, for example, serve as media of exchange, and they are a debt to the banks that issue them. Privately-issued banknotes can serve as hand-to-hand currency. They serve as media of exchange, but they are credit money. They are debt to the issuing banks.

From the point of view of the firms and households making payments, these bank liabilities are assets that serve as media of exchange. They can quote prices in terms of them, receive payments of them, and then make payments with them.

Now, if the monetary liabilities of many banks are to be accepted at par, there needs to be some kind of clearing system between the banks. It is possible to have a settlement system where the quantity of the settlement medium demanded in equilibrium is zero. And so, it is arguable that there is no base money. All the money is credit money.

However, even if banks do hold balances of some settlement medium, that doesn't mean that it is necessarily "outside money," that is, a liability of no one. Suppose inter-bank clearings are handled by a private clearinghouse, and the clearinghouse creates (and destroys) balances by ordinary open market operations. It purchases and sells assets. The balances are liabilities to the clearinghouse. It is another type of debt.

Whether or not it is desirable, a pure credit money system is thinkable. There is money. Prices can be quoted in terms of that money. A price level can be calculated in terms of that money. It is even conceivable that the quantity of reserve balances created by a private clearinghouse would be adjusted to stabilize some other nominal value, like a growth path for nominal GDP. For example, if the clearinghouse was obligated to maintain index futures convertibility.

Now, if we consider the status quo, where an independent government agency issues base money, the only real question is whether the currency and reserve balances are best understood as a special type of debt or not. This very much depends on how serious is the commitment to the target for some nominal variable like inflation or a growth path of nominal GDP. A monetary authority organized on banking principles and subject to index future convertibility, using open market operations to change the quantity of base money, and so matching its issue of base money with financial assets, looks very similar to the private clearinghouse described above.

Is the current Fed's commitment to its inflation target enough? It is an empirical question. And Ashwin Parameswaran was correct about that.